locally chaos cosmos

Preview

During Public Preview you can use the Team plan for free, and get an early-adopter discount at GA. Sign in

locally chaos cosmos

This page is part of the CLI Reference and is generated from the CLI's own help output, so it matches what locally chaos cosmos --help prints.

For more information see Chaos Engineering and Cosmos DB Emulator.

locally chaos cosmos

Simulate Cosmos DB (data plane) failure modes.

Simulate Cosmos DB (data plane) failure modes.

A fault command sets the fault's rate. With a scope flag (--location / --name)
it adds or edits a rule instead.

Examples:
locally chaos cosmos <fault> --percentage N
  locally chaos cosmos <fault> reset
  locally chaos cosmos show
  locally chaos cosmos reset

Usage

locally chaos cosmos [flags]
  locally chaos cosmos [command]

Flags

-h, --help[=true|false]   help for cosmos

Subcommands

locally chaos cosmos 403

Answer requests with 403, as when a role assignment has not replicated yet.

Answer requests with 403, as when a role assignment has not replicated yet. On
the control plane the error code is AuthorizationFailed; on a database the
engine's own equivalent error is sent instead.

Without scope flags this sets the fault's rate on Cosmos DB (data plane). With
--location / --name it adds or edits one rule, leaving the rate and other rules
alone.

A rule can be lower than the rate: "everywhere except one place" is a rate of
100 plus a rule at 0.

--name acct1 matches that resource in any resource group; --name rg1/acct1
matches only rg1's.

Examples:
locally chaos cosmos 403 --percentage 25
  locally chaos cosmos 403 --percentage 100 --location london
  locally chaos cosmos 403 reset --location london
  locally chaos cosmos 403 --percentage 100 --name acct1

Usage

locally chaos cosmos 403 [flags]
  locally chaos cosmos 403 [command]

Flags

-h, --help[=true|false]     help for 403
      --location string       Scope this rule to one location, e.g. london
      --name string           Scope this rule to one resource by name, e.g. acct1, or rg1/acct1 for one resource group's
      --percentage int        How often this fault fires, 0-100 (required)
      --subscription string   Subscription UUID (defaults to the configured default)

Subcommands

locally chaos cosmos 403 reset

Remove one 403 rule, or put the whole fault back to zero.

Reset the 403 fault.

With a scope flag, removes that rule so those requests use the fault's rate.
Without one, removes every rule and sets the rate to 0.

'reset --location X' removes the rule; '--percentage 0 --location X' keeps a
rule that never fires, which is how to say "everywhere except X".

Usage

locally chaos cosmos 403 reset [flags]

Flags

-h, --help[=true|false]     help for reset
      --location string       Scope this rule to one location, e.g. london
      --name string           Scope this rule to one resource by name, e.g. acct1, or rg1/acct1 for one resource group's
      --subscription string   Subscription UUID (defaults to the configured default)

locally chaos cosmos 429

Answer requests with 429 Too Many Requests and a retry hint.

Answer requests with 429 Too Many Requests and a retry hint. On a database the
engine's own throttling error is sent instead.

Without scope flags this sets the fault's rate on Cosmos DB (data plane). With
--location / --name it adds or edits one rule, leaving the rate and other rules
alone.

A rule can be lower than the rate: "everywhere except one place" is a rate of
100 plus a rule at 0.

--name acct1 matches that resource in any resource group; --name rg1/acct1
matches only rg1's.

--retry-after sets the Retry-After header the fault answers with, such as 5s.
Without one the service's usual hint is sent.

A value flag left out keeps what the row already has; a new rule starts with
none. 'reset' clears a row.

Examples:
locally chaos cosmos 429 --percentage 25
  locally chaos cosmos 429 --percentage 25 --retry-after 5s
  locally chaos cosmos 429 --percentage 100 --location london
  locally chaos cosmos 429 reset --location london
  locally chaos cosmos 429 --percentage 100 --name acct1

Usage

locally chaos cosmos 429 [flags]
  locally chaos cosmos 429 [command]

Flags

-h, --help[=true|false]     help for 429
      --location string       Scope this rule to one location, e.g. london
      --name string           Scope this rule to one resource by name, e.g. acct1, or rg1/acct1 for one resource group's
      --percentage int        How often this fault fires, 0-100 (required)
      --retry-after string    Retry-After this fault answers with, e.g. 5s or 500ms (the service's usual hint when omitted)
      --subscription string   Subscription UUID (defaults to the configured default)

Subcommands

locally chaos cosmos 429 reset

Remove one 429 rule, or put the whole fault back to zero.

Reset the 429 fault.

With a scope flag, removes that rule so those requests use the fault's rate.
Without one, removes every rule and sets the rate to 0.

'reset --location X' removes the rule; '--percentage 0 --location X' keeps a
rule that never fires, which is how to say "everywhere except X".

Usage

locally chaos cosmos 429 reset [flags]

Flags

-h, --help[=true|false]     help for reset
      --location string       Scope this rule to one location, e.g. london
      --name string           Scope this rule to one resource by name, e.g. acct1, or rg1/acct1 for one resource group's
      --subscription string   Subscription UUID (defaults to the configured default)

locally chaos cosmos 500

Answer requests with 500 Internal Server Error.

Answer requests with 500 Internal Server Error. On a database the engine's own
equivalent error is sent instead.

Without scope flags this sets the fault's rate on Cosmos DB (data plane). With
--location / --name it adds or edits one rule, leaving the rate and other rules
alone.

A rule can be lower than the rate: "everywhere except one place" is a rate of
100 plus a rule at 0.

--name acct1 matches that resource in any resource group; --name rg1/acct1
matches only rg1's.

Examples:
locally chaos cosmos 500 --percentage 25
  locally chaos cosmos 500 --percentage 100 --location london
  locally chaos cosmos 500 reset --location london
  locally chaos cosmos 500 --percentage 100 --name acct1

Usage

locally chaos cosmos 500 [flags]
  locally chaos cosmos 500 [command]

Flags

-h, --help[=true|false]     help for 500
      --location string       Scope this rule to one location, e.g. london
      --name string           Scope this rule to one resource by name, e.g. acct1, or rg1/acct1 for one resource group's
      --percentage int        How often this fault fires, 0-100 (required)
      --subscription string   Subscription UUID (defaults to the configured default)

Subcommands

locally chaos cosmos 500 reset

Remove one 500 rule, or put the whole fault back to zero.

Reset the 500 fault.

With a scope flag, removes that rule so those requests use the fault's rate.
Without one, removes every rule and sets the rate to 0.

'reset --location X' removes the rule; '--percentage 0 --location X' keeps a
rule that never fires, which is how to say "everywhere except X".

Usage

locally chaos cosmos 500 reset [flags]

Flags

-h, --help[=true|false]     help for reset
      --location string       Scope this rule to one location, e.g. london
      --name string           Scope this rule to one resource by name, e.g. acct1, or rg1/acct1 for one resource group's
      --subscription string   Subscription UUID (defaults to the configured default)

locally chaos cosmos disconnect

Close live connections cleanly, in the protocol's own terms, so the client has to reconnect.

Close live connections cleanly, in the protocol's own terms, so the client has
to reconnect. Every interval each open connection is rolled for. Compare
drop-connection, which cuts them with nothing said.

Without scope flags this sets the fault's rate on Cosmos DB (data plane). With
--location / --name it adds or edits one rule, leaving the rate and other rules
alone.

A rule can be lower than the rate: "everywhere except one place" is a rate of
100 plus a rule at 0.

--name acct1 matches that resource in any resource group; --name rg1/acct1
matches only rg1's.

--interval is how often the fault starts, such as 10m. It is required whenever
the fault fires: nothing else sets it off.

A value flag left out keeps what the row already has; a new rule starts with
none. 'reset' clears a row.

Examples:
locally chaos cosmos disconnect --percentage 25 --interval 10m
  locally chaos cosmos disconnect --percentage 100 --interval 10m --location london
  locally chaos cosmos disconnect reset --location london
  locally chaos cosmos disconnect --percentage 100 --interval 10m --name acct1

Usage

locally chaos cosmos disconnect [flags]
  locally chaos cosmos disconnect [command]

Flags

-h, --help[=true|false]     help for disconnect
      --interval string       How often this fault rolls for each resource, e.g. 10m; the first roll lands at a random point within the first interval (required when it fires)
      --location string       Scope this rule to one location, e.g. london
      --name string           Scope this rule to one resource by name, e.g. acct1, or rg1/acct1 for one resource group's
      --percentage int        How often this fault fires, 0-100 (required)
      --subscription string   Subscription UUID (defaults to the configured default)

Subcommands

locally chaos cosmos disconnect reset

Remove one disconnect rule, or put the whole fault back to zero.

Reset the disconnect fault.

With a scope flag, removes that rule so those requests use the fault's rate.
Without one, removes every rule and sets the rate to 0.

'reset --location X' removes the rule; '--percentage 0 --location X' keeps a
rule that never fires, which is how to say "everywhere except X".

Usage

locally chaos cosmos disconnect reset [flags]

Flags

-h, --help[=true|false]     help for reset
      --location string       Scope this rule to one location, e.g. london
      --name string           Scope this rule to one resource by name, e.g. acct1, or rg1/acct1 for one resource group's
      --subscription string   Subscription UUID (defaults to the configured default)

locally chaos cosmos drop-connection

Close the connection without answering.

Close the connection without answering. The client sees the connection reset,
with no status code or error body. With --interval it also cuts live
connections: every interval each is rolled for, and on a hit it is reset with
nothing said - no close frame or goodbye. For a clean close, use disconnect.

Without scope flags this sets the fault's rate on Cosmos DB (data plane). With
--location / --name it adds or edits one rule, leaving the rate and other rules
alone.

A rule can be lower than the rate: "everywhere except one place" is a rate of
100 plus a rule at 0.

--name acct1 matches that resource in any resource group; --name rg1/acct1
matches only rg1's.

--interval adds periodic action on live connections, such as every 30s, to the
fault's per-request behaviour, which carries on as before. For periodic cuts
with no per-request drops, arm disconnect instead.

A value flag left out keeps what the row already has; a new rule starts with
none. 'reset' clears a row.

Examples:
locally chaos cosmos drop-connection --percentage 25
  locally chaos cosmos drop-connection --percentage 25 --interval 30s
  locally chaos cosmos drop-connection --percentage 100 --location london
  locally chaos cosmos drop-connection reset --location london
  locally chaos cosmos drop-connection --percentage 100 --name acct1

Usage

locally chaos cosmos drop-connection [flags]
  locally chaos cosmos drop-connection [command]

Flags

-h, --help[=true|false]     help for drop-connection
      --interval string       Also act on live connections this often, e.g. 30s or 10m (adds to the per-request fault; use disconnect for periodic cuts only)
      --location string       Scope this rule to one location, e.g. london
      --name string           Scope this rule to one resource by name, e.g. acct1, or rg1/acct1 for one resource group's
      --percentage int        How often this fault fires, 0-100 (required)
      --subscription string   Subscription UUID (defaults to the configured default)

Subcommands

locally chaos cosmos drop-connection reset

Remove one drop-connection rule, or put the whole fault back to zero.

Reset the drop-connection fault.

With a scope flag, removes that rule so those requests use the fault's rate.
Without one, removes every rule and sets the rate to 0.

'reset --location X' removes the rule; '--percentage 0 --location X' keeps a
rule that never fires, which is how to say "everywhere except X".

Usage

locally chaos cosmos drop-connection reset [flags]

Flags

-h, --help[=true|false]     help for reset
      --location string       Scope this rule to one location, e.g. london
      --name string           Scope this rule to one resource by name, e.g. acct1, or rg1/acct1 for one resource group's
      --subscription string   Subscription UUID (defaults to the configured default)

locally chaos cosmos failover

Fail the primary over on a schedule.

Fail the primary over on a schedule. On a hit the primary is unavailable for the
rolled duration and any secondary serves reads only, then the primary returns.

Without scope flags this sets the fault's rate on Cosmos DB (data plane). With
--location / --name it adds or edits one rule, leaving the rate and other rules
alone.

A rule can be lower than the rate: "everywhere except one place" is a rate of
100 plus a rule at 0.

--name acct1 matches that resource in any resource group; --name rg1/acct1
matches only rg1's.

--seconds and --millis set the duration. A bare number is in the flag's unit;
one with its own suffix, such as 250ms or 1.5s, means what it says. One value N
picks 0 to N each time; MIN-MAX, such as 30-120, picks within that range. A
duration is required whenever the fault fires.

--interval is how often the fault starts, such as 10m. It is required whenever
the fault fires: nothing else sets it off.

A value flag left out keeps what the row already has; a new rule starts with
none. 'reset' clears a row.

Examples:
locally chaos cosmos failover --percentage 25 --seconds 30-120 --interval 10m
  locally chaos cosmos failover --percentage 100 --seconds 30-120 --interval 10m --location london
  locally chaos cosmos failover reset --location london
  locally chaos cosmos failover --percentage 100 --seconds 30-120 --interval 10m --name acct1

Usage

locally chaos cosmos failover [flags]
  locally chaos cosmos failover [command]

Flags

-h, --help[=true|false]     help for failover
      --interval string       How often this fault rolls for each resource, e.g. 10m; the first roll lands at a random point within the first interval (required when it fires)
      --location string       Scope this rule to one location, e.g. london
      --millis string         Duration in milliseconds, e.g. 250; N for 0 to N, or MIN-MAX for a range
      --name string           Scope this rule to one resource by name, e.g. acct1, or rg1/acct1 for one resource group's
      --percentage int        How often this fault fires, 0-100 (required)
      --seconds string        Duration in seconds, e.g. 30 or 1.5; N for 0 to N, or MIN-MAX for a range
      --subscription string   Subscription UUID (defaults to the configured default)

Subcommands

locally chaos cosmos failover reset

Remove one failover rule, or put the whole fault back to zero.

Reset the failover fault.

With a scope flag, removes that rule so those requests use the fault's rate.
Without one, removes every rule and sets the rate to 0.

'reset --location X' removes the rule; '--percentage 0 --location X' keeps a
rule that never fires, which is how to say "everywhere except X".

Usage

locally chaos cosmos failover reset [flags]

Flags

-h, --help[=true|false]     help for reset
      --location string       Scope this rule to one location, e.g. london
      --name string           Scope this rule to one resource by name, e.g. acct1, or rg1/acct1 for one resource group's
      --subscription string   Subscription UUID (defaults to the configured default)

locally chaos cosmos incident

Take each resource down for a while, then bring it back.

Take each resource down for a while, then bring it back. Every interval each
resource (or, on the directory and identity surfaces, the tenant) rolls against
the percentage; on a hit every request to it fails for the rolled duration,
answered as an outage would be, then it recovers on its own. Compare outage,
which fails requests for as long as it is armed.

Without scope flags this sets the fault's rate on Cosmos DB (data plane). With
--location / --name it adds or edits one rule, leaving the rate and other rules
alone.

A rule can be lower than the rate: "everywhere except one place" is a rate of
100 plus a rule at 0.

--name acct1 matches that resource in any resource group; --name rg1/acct1
matches only rg1's.

--seconds and --millis set the duration. A bare number is in the flag's unit;
one with its own suffix, such as 250ms or 1.5s, means what it says. One value N
picks 0 to N each time; MIN-MAX, such as 30-120, picks within that range. A
duration is required whenever the fault fires.

--interval is how often the fault starts, such as 10m. It is required whenever
the fault fires: nothing else sets it off.

A value flag left out keeps what the row already has; a new rule starts with
none. 'reset' clears a row.

Examples:
locally chaos cosmos incident --percentage 25 --seconds 30-120 --interval 10m
  locally chaos cosmos incident --percentage 100 --seconds 30-120 --interval 10m --location london
  locally chaos cosmos incident reset --location london
  locally chaos cosmos incident --percentage 100 --seconds 30-120 --interval 10m --name acct1

Usage

locally chaos cosmos incident [flags]
  locally chaos cosmos incident [command]

Flags

-h, --help[=true|false]     help for incident
      --interval string       How often this fault rolls for each resource, e.g. 10m; the first roll lands at a random point within the first interval (required when it fires)
      --location string       Scope this rule to one location, e.g. london
      --millis string         Duration in milliseconds, e.g. 250; N for 0 to N, or MIN-MAX for a range
      --name string           Scope this rule to one resource by name, e.g. acct1, or rg1/acct1 for one resource group's
      --percentage int        How often this fault fires, 0-100 (required)
      --seconds string        Duration in seconds, e.g. 30 or 1.5; N for 0 to N, or MIN-MAX for a range
      --subscription string   Subscription UUID (defaults to the configured default)

Subcommands

locally chaos cosmos incident reset

Remove one incident rule, or put the whole fault back to zero.

Reset the incident fault.

With a scope flag, removes that rule so those requests use the fault's rate.
Without one, removes every rule and sets the rate to 0.

'reset --location X' removes the rule; '--percentage 0 --location X' keeps a
rule that never fires, which is how to say "everywhere except X".

Usage

locally chaos cosmos incident reset [flags]

Flags

-h, --help[=true|false]     help for reset
      --location string       Scope this rule to one location, e.g. london
      --name string           Scope this rule to one resource by name, e.g. acct1, or rg1/acct1 for one resource group's
      --subscription string   Subscription UUID (defaults to the configured default)

locally chaos cosmos latency

Hold requests for a random delay before they are handled.

Hold requests for a random delay before they are handled. The response is
unchanged, just late. With --interval it also stalls live connections: every
interval each is rolled for, and on a hit nothing crosses it in either direction
until the rolled delay has passed. Nothing is dropped; it all arrives late.

Without scope flags this sets the fault's rate on Cosmos DB (data plane). With
--location / --name it adds or edits one rule, leaving the rate and other rules
alone.

A rule can be lower than the rate: "everywhere except one place" is a rate of
100 plus a rule at 0.

--name acct1 matches that resource in any resource group; --name rg1/acct1
matches only rg1's.

--seconds and --millis set the duration. A bare number is in the flag's unit;
one with its own suffix, such as 250ms or 1.5s, means what it says. The value is
a maximum: each firing picks 0 to N. A duration is required whenever the fault
fires.

--interval adds periodic action on live connections, such as every 30s, to the
fault's per-request behaviour, which carries on as before. For periodic cuts
with no per-request drops, arm disconnect instead.

A value flag left out keeps what the row already has; a new rule starts with
none. 'reset' clears a row.

Examples:
locally chaos cosmos latency --percentage 25 --seconds 30
  locally chaos cosmos latency --percentage 25 --seconds 30 --interval 30s
  locally chaos cosmos latency --percentage 100 --seconds 30 --location london
  locally chaos cosmos latency reset --location london
  locally chaos cosmos latency --percentage 100 --seconds 30 --name acct1

Usage

locally chaos cosmos latency [flags]
  locally chaos cosmos latency [command]

Flags

-h, --help[=true|false]     help for latency
      --interval string       Also act on live connections this often, e.g. 30s or 10m (adds to the per-request fault; use disconnect for periodic cuts only)
      --location string       Scope this rule to one location, e.g. london
      --millis string         Duration in milliseconds, e.g. 250; the maximum: each firing picks 0 to N
      --name string           Scope this rule to one resource by name, e.g. acct1, or rg1/acct1 for one resource group's
      --percentage int        How often this fault fires, 0-100 (required)
      --seconds string        Duration in seconds, e.g. 30 or 1.5; the maximum: each firing picks 0 to N
      --subscription string   Subscription UUID (defaults to the configured default)

Subcommands

locally chaos cosmos latency reset

Remove one latency rule, or put the whole fault back to zero.

Reset the latency fault.

With a scope flag, removes that rule so those requests use the fault's rate.
Without one, removes every rule and sets the rate to 0.

'reset --location X' removes the rule; '--percentage 0 --location X' keeps a
rule that never fires, which is how to say "everywhere except X".

Usage

locally chaos cosmos latency reset [flags]

Flags

-h, --help[=true|false]     help for reset
      --location string       Scope this rule to one location, e.g. london
      --name string           Scope this rule to one resource by name, e.g. acct1, or rg1/acct1 for one resource group's
      --subscription string   Subscription UUID (defaults to the configured default)

locally chaos cosmos outage

Fail requests for as long as the rule is armed, like a region or service that is down.

Fail requests for as long as the rule is armed, like a region or service that is
down. Each request rolls against the percentage, so at 100 every request fails
until the fault is turned off. Each service answers in its own terms, usually
503 Service Unavailable, and a protocol with live connections also closes the
ones open when the outage starts. For downtime that ends by itself, see
incident.

Without scope flags this sets the fault's rate on Cosmos DB (data plane). With
--location / --name it adds or edits one rule, leaving the rate and other rules
alone.

A rule can be lower than the rate: "everywhere except one place" is a rate of
100 plus a rule at 0.

--name acct1 matches that resource in any resource group; --name rg1/acct1
matches only rg1's.

--retry-after sets the Retry-After header the fault answers with, such as 5s.
Without one the service's usual hint is sent.

A value flag left out keeps what the row already has; a new rule starts with
none. 'reset' clears a row.

Examples:
locally chaos cosmos outage --percentage 25
  locally chaos cosmos outage --percentage 25 --retry-after 5s
  locally chaos cosmos outage --percentage 100 --location london
  locally chaos cosmos outage reset --location london
  locally chaos cosmos outage --percentage 100 --name acct1

Usage

locally chaos cosmos outage [flags]
  locally chaos cosmos outage [command]

Flags

-h, --help[=true|false]     help for outage
      --location string       Scope this rule to one location, e.g. london
      --name string           Scope this rule to one resource by name, e.g. acct1, or rg1/acct1 for one resource group's
      --percentage int        How often this fault fires, 0-100 (required)
      --retry-after string    Retry-After this fault answers with, e.g. 5s or 500ms (the service's usual hint when omitted)
      --subscription string   Subscription UUID (defaults to the configured default)

Subcommands

locally chaos cosmos outage reset

Remove one outage rule, or put the whole fault back to zero.

Reset the outage fault.

With a scope flag, removes that rule so those requests use the fault's rate.
Without one, removes every rule and sets the rate to 0.

'reset --location X' removes the rule; '--percentage 0 --location X' keeps a
rule that never fires, which is how to say "everywhere except X".

Usage

locally chaos cosmos outage reset [flags]

Flags

-h, --help[=true|false]     help for reset
      --location string       Scope this rule to one location, e.g. london
      --name string           Scope this rule to one resource by name, e.g. acct1, or rg1/acct1 for one resource group's
      --subscription string   Subscription UUID (defaults to the configured default)

locally chaos cosmos partial-response

Send the status, headers and part of the body, then cut the connection.

Send the status, headers and part of the body, then cut the connection. The
client sees a response that ends early.

Without scope flags this sets the fault's rate on Cosmos DB (data plane). With
--location / --name it adds or edits one rule, leaving the rate and other rules
alone.

A rule can be lower than the rate: "everywhere except one place" is a rate of
100 plus a rule at 0.

--name acct1 matches that resource in any resource group; --name rg1/acct1
matches only rg1's.

Examples:
locally chaos cosmos partial-response --percentage 25
  locally chaos cosmos partial-response --percentage 100 --location london
  locally chaos cosmos partial-response reset --location london
  locally chaos cosmos partial-response --percentage 100 --name acct1

Usage

locally chaos cosmos partial-response [flags]
  locally chaos cosmos partial-response [command]

Flags

-h, --help[=true|false]     help for partial-response
      --location string       Scope this rule to one location, e.g. london
      --name string           Scope this rule to one resource by name, e.g. acct1, or rg1/acct1 for one resource group's
      --percentage int        How often this fault fires, 0-100 (required)
      --subscription string   Subscription UUID (defaults to the configured default)

Subcommands

locally chaos cosmos partial-response reset

Remove one partial-response rule, or put the whole fault back to zero.

Reset the partial-response fault.

With a scope flag, removes that rule so those requests use the fault's rate.
Without one, removes every rule and sets the rate to 0.

'reset --location X' removes the rule; '--percentage 0 --location X' keeps a
rule that never fires, which is how to say "everywhere except X".

Usage

locally chaos cosmos partial-response reset [flags]

Flags

-h, --help[=true|false]     help for reset
      --location string       Scope this rule to one location, e.g. london
      --name string           Scope this rule to one resource by name, e.g. acct1, or rg1/acct1 for one resource group's
      --subscription string   Subscription UUID (defaults to the configured default)

locally chaos cosmos partition-split

Answer requests with 410 Gone substatus 1002, as Cosmos DB does while a partition splits.

Answer requests with 410 Gone substatus 1002, as Cosmos DB does while a
partition splits. The client is expected to refresh its routing and retry.

Without scope flags this sets the fault's rate on Cosmos DB (data plane). With
--location / --name it adds or edits one rule, leaving the rate and other rules
alone.

A rule can be lower than the rate: "everywhere except one place" is a rate of
100 plus a rule at 0.

--name acct1 matches that resource in any resource group; --name rg1/acct1
matches only rg1's.

Examples:
locally chaos cosmos partition-split --percentage 25
  locally chaos cosmos partition-split --percentage 100 --location london
  locally chaos cosmos partition-split reset --location london
  locally chaos cosmos partition-split --percentage 100 --name acct1

Usage

locally chaos cosmos partition-split [flags]
  locally chaos cosmos partition-split [command]

Flags

-h, --help[=true|false]     help for partition-split
      --location string       Scope this rule to one location, e.g. london
      --name string           Scope this rule to one resource by name, e.g. acct1, or rg1/acct1 for one resource group's
      --percentage int        How often this fault fires, 0-100 (required)
      --subscription string   Subscription UUID (defaults to the configured default)

Subcommands

locally chaos cosmos partition-split reset

Remove one partition-split rule, or put the whole fault back to zero.

Reset the partition-split fault.

With a scope flag, removes that rule so those requests use the fault's rate.
Without one, removes every rule and sets the rate to 0.

'reset --location X' removes the rule; '--percentage 0 --location X' keeps a
rule that never fires, which is how to say "everywhere except X".

Usage

locally chaos cosmos partition-split reset [flags]

Flags

-h, --help[=true|false]     help for reset
      --location string       Scope this rule to one location, e.g. london
      --name string           Scope this rule to one resource by name, e.g. acct1, or rg1/acct1 for one resource group's
      --subscription string   Subscription UUID (defaults to the configured default)

locally chaos cosmos request-timeout

Fail requests with the service's own server-side timeout, such as Cosmos DB's 408 or Storage's OperationTimedOut.

Fail requests with the service's own server-side timeout, such as Cosmos DB's
408 or Storage's OperationTimedOut.

Without scope flags this sets the fault's rate on Cosmos DB (data plane). With
--location / --name it adds or edits one rule, leaving the rate and other rules
alone.

A rule can be lower than the rate: "everywhere except one place" is a rate of
100 plus a rule at 0.

--name acct1 matches that resource in any resource group; --name rg1/acct1
matches only rg1's.

Examples:
locally chaos cosmos request-timeout --percentage 25
  locally chaos cosmos request-timeout --percentage 100 --location london
  locally chaos cosmos request-timeout reset --location london
  locally chaos cosmos request-timeout --percentage 100 --name acct1

Usage

locally chaos cosmos request-timeout [flags]
  locally chaos cosmos request-timeout [command]

Flags

-h, --help[=true|false]     help for request-timeout
      --location string       Scope this rule to one location, e.g. london
      --name string           Scope this rule to one resource by name, e.g. acct1, or rg1/acct1 for one resource group's
      --percentage int        How often this fault fires, 0-100 (required)
      --subscription string   Subscription UUID (defaults to the configured default)

Subcommands

locally chaos cosmos request-timeout reset

Remove one request-timeout rule, or put the whole fault back to zero.

Reset the request-timeout fault.

With a scope flag, removes that rule so those requests use the fault's rate.
Without one, removes every rule and sets the rate to 0.

'reset --location X' removes the rule; '--percentage 0 --location X' keeps a
rule that never fires, which is how to say "everywhere except X".

Usage

locally chaos cosmos request-timeout reset [flags]

Flags

-h, --help[=true|false]     help for reset
      --location string       Scope this rule to one location, e.g. london
      --name string           Scope this rule to one resource by name, e.g. acct1, or rg1/acct1 for one resource group's
      --subscription string   Subscription UUID (defaults to the configured default)

locally chaos cosmos reset

Zero every Cosmos DB (data plane) fault and drop every rule.

Zero every Cosmos DB (data plane) fault and drop every rule.

Examples:
locally chaos cosmos reset

Usage

locally chaos cosmos reset [flags]

Flags

-h, --help[=true|false]     help for reset
      --subscription string   Subscription UUID (defaults to the configured default)

locally chaos cosmos retry-with

Answer writes with 449 Retry With, as Cosmos DB does on a write conflict it expects the client to retry.

Answer writes with 449 Retry With, as Cosmos DB does on a write conflict it
expects the client to retry.

Without scope flags this sets the fault's rate on Cosmos DB (data plane). With
--location / --name it adds or edits one rule, leaving the rate and other rules
alone.

A rule can be lower than the rate: "everywhere except one place" is a rate of
100 plus a rule at 0.

--name acct1 matches that resource in any resource group; --name rg1/acct1
matches only rg1's.

Examples:
locally chaos cosmos retry-with --percentage 25
  locally chaos cosmos retry-with --percentage 100 --location london
  locally chaos cosmos retry-with reset --location london
  locally chaos cosmos retry-with --percentage 100 --name acct1

Usage

locally chaos cosmos retry-with [flags]
  locally chaos cosmos retry-with [command]

Flags

-h, --help[=true|false]     help for retry-with
      --location string       Scope this rule to one location, e.g. london
      --name string           Scope this rule to one resource by name, e.g. acct1, or rg1/acct1 for one resource group's
      --percentage int        How often this fault fires, 0-100 (required)
      --subscription string   Subscription UUID (defaults to the configured default)

Subcommands

locally chaos cosmos retry-with reset

Remove one retry-with rule, or put the whole fault back to zero.

Reset the retry-with fault.

With a scope flag, removes that rule so those requests use the fault's rate.
Without one, removes every rule and sets the rate to 0.

'reset --location X' removes the rule; '--percentage 0 --location X' keeps a
rule that never fires, which is how to say "everywhere except X".

Usage

locally chaos cosmos retry-with reset [flags]

Flags

-h, --help[=true|false]     help for reset
      --location string       Scope this rule to one location, e.g. london
      --name string           Scope this rule to one resource by name, e.g. acct1, or rg1/acct1 for one resource group's
      --subscription string   Subscription UUID (defaults to the configured default)

locally chaos cosmos session-not-available

Answer reads with 404 substatus 1002, as Cosmos DB does when a replica has not caught up with the client's session.

Answer reads with 404 substatus 1002, as Cosmos DB does when a replica has not
caught up with the client's session.

Without scope flags this sets the fault's rate on Cosmos DB (data plane). With
--location / --name it adds or edits one rule, leaving the rate and other rules
alone.

A rule can be lower than the rate: "everywhere except one place" is a rate of
100 plus a rule at 0.

--name acct1 matches that resource in any resource group; --name rg1/acct1
matches only rg1's.

Examples:
locally chaos cosmos session-not-available --percentage 25
  locally chaos cosmos session-not-available --percentage 100 --location london
  locally chaos cosmos session-not-available reset --location london
  locally chaos cosmos session-not-available --percentage 100 --name acct1

Usage

locally chaos cosmos session-not-available [flags]
  locally chaos cosmos session-not-available [command]

Flags

-h, --help[=true|false]     help for session-not-available
      --location string       Scope this rule to one location, e.g. london
      --name string           Scope this rule to one resource by name, e.g. acct1, or rg1/acct1 for one resource group's
      --percentage int        How often this fault fires, 0-100 (required)
      --subscription string   Subscription UUID (defaults to the configured default)

Subcommands

locally chaos cosmos session-not-available reset

Remove one session-not-available rule, or put the whole fault back to zero.

Reset the session-not-available fault.

With a scope flag, removes that rule so those requests use the fault's rate.
Without one, removes every rule and sets the rate to 0.

'reset --location X' removes the rule; '--percentage 0 --location X' keeps a
rule that never fires, which is how to say "everywhere except X".

Usage

locally chaos cosmos session-not-available reset [flags]

Flags

-h, --help[=true|false]     help for reset
      --location string       Scope this rule to one location, e.g. london
      --name string           Scope this rule to one resource by name, e.g. acct1, or rg1/acct1 for one resource group's
      --subscription string   Subscription UUID (defaults to the configured default)

locally chaos cosmos show

Print the current Cosmos DB (data plane) chaos configuration.

Print the current Cosmos DB (data plane) chaos configuration.

Examples:
locally chaos cosmos show

Usage

locally chaos cosmos show [flags]

Flags

-h, --help[=true|false]     help for show
      --subscription string   Subscription UUID (defaults to the configured default)

locally chaos cosmos stale-read

Serve stale reads after a write.

Serve stale reads after a write. Each write opens a window on what it wrote, and
reads inside the window see it as it was before the write.

Without scope flags this sets the fault's rate on Cosmos DB (data plane). With
--location / --name it adds or edits one rule, leaving the rate and other rules
alone.

A rule can be lower than the rate: "everywhere except one place" is a rate of
100 plus a rule at 0.

--name acct1 matches that resource in any resource group; --name rg1/acct1
matches only rg1's.

--seconds and --millis set the duration. A bare number is in the flag's unit;
one with its own suffix, such as 250ms or 1.5s, means what it says. The value is
a maximum: each firing picks 0 to N. A duration is required whenever the fault
fires.

A value flag left out keeps what the row already has; a new rule starts with
none. 'reset' clears a row.

Examples:
locally chaos cosmos stale-read --percentage 25 --seconds 30
  locally chaos cosmos stale-read --percentage 100 --seconds 30 --location london
  locally chaos cosmos stale-read reset --location london
  locally chaos cosmos stale-read --percentage 100 --seconds 30 --name acct1

Usage

locally chaos cosmos stale-read [flags]
  locally chaos cosmos stale-read [command]

Flags

-h, --help[=true|false]     help for stale-read
      --location string       Scope this rule to one location, e.g. london
      --millis string         Duration in milliseconds, e.g. 250; the maximum: each firing picks 0 to N
      --name string           Scope this rule to one resource by name, e.g. acct1, or rg1/acct1 for one resource group's
      --percentage int        How often this fault fires, 0-100 (required)
      --seconds string        Duration in seconds, e.g. 30 or 1.5; the maximum: each firing picks 0 to N
      --subscription string   Subscription UUID (defaults to the configured default)

Subcommands

locally chaos cosmos stale-read reset

Remove one stale-read rule, or put the whole fault back to zero.

Reset the stale-read fault.

With a scope flag, removes that rule so those requests use the fault's rate.
Without one, removes every rule and sets the rate to 0.

'reset --location X' removes the rule; '--percentage 0 --location X' keeps a
rule that never fires, which is how to say "everywhere except X".

Usage

locally chaos cosmos stale-read reset [flags]

Flags

-h, --help[=true|false]     help for reset
      --location string       Scope this rule to one location, e.g. london
      --name string           Scope this rule to one resource by name, e.g. acct1, or rg1/acct1 for one resource group's
      --subscription string   Subscription UUID (defaults to the configured default)

locally chaos cosmos write-forbidden

Refuse writes with 403 substatus 3, as a Cosmos DB region that is not accepting writes does.

Refuse writes with 403 substatus 3, as a Cosmos DB region that is not accepting
writes does.

Without scope flags this sets the fault's rate on Cosmos DB (data plane). With
--location / --name it adds or edits one rule, leaving the rate and other rules
alone.

A rule can be lower than the rate: "everywhere except one place" is a rate of
100 plus a rule at 0.

--name acct1 matches that resource in any resource group; --name rg1/acct1
matches only rg1's.

Examples:
locally chaos cosmos write-forbidden --percentage 25
  locally chaos cosmos write-forbidden --percentage 100 --location london
  locally chaos cosmos write-forbidden reset --location london
  locally chaos cosmos write-forbidden --percentage 100 --name acct1

Usage

locally chaos cosmos write-forbidden [flags]
  locally chaos cosmos write-forbidden [command]

Flags

-h, --help[=true|false]     help for write-forbidden
      --location string       Scope this rule to one location, e.g. london
      --name string           Scope this rule to one resource by name, e.g. acct1, or rg1/acct1 for one resource group's
      --percentage int        How often this fault fires, 0-100 (required)
      --subscription string   Subscription UUID (defaults to the configured default)

Subcommands

locally chaos cosmos write-forbidden reset

Remove one write-forbidden rule, or put the whole fault back to zero.

Reset the write-forbidden fault.

With a scope flag, removes that rule so those requests use the fault's rate.
Without one, removes every rule and sets the rate to 0.

'reset --location X' removes the rule; '--percentage 0 --location X' keeps a
rule that never fires, which is how to say "everywhere except X".

Usage

locally chaos cosmos write-forbidden reset [flags]

Flags

-h, --help[=true|false]     help for reset
      --location string       Scope this rule to one location, e.g. london
      --name string           Scope this rule to one resource by name, e.g. acct1, or rg1/acct1 for one resource group's
      --subscription string   Subscription UUID (defaults to the configured default)
Preview

During Public Preview you can use the Team plan for free, and get an early-adopter discount at GA. Sign in

A local cloud for you and your AI agents.

Your Azure infrastructure, running on your machine. Deploy in seconds, break things freely, and ship to Azure when you're ready.