package limit

Get desktop application:
View/edit binary Protocol Buffers messages

• rpc ClearBindings (LimitClearBindingsRequest, LimitClearBindingsResponse)

  limit.proto:61

  Remove bindings from a limit, scoped by origin. The escape hatch for a mis-set exit status; SPEC bindings are never removable this way.

  message LimitClearBindingsRequest

  limit.proto:353

  ClearBindings

  • string limit_name = 1

  • repeated LimitBindSource sources = 2

    Only AUTO and MANUAL are accepted; SPEC bindings are never cleared here.

  message LimitClearBindingsResponse

  limit.proto:359

  • int32 removed = 1

• rpc Create (LimitCreateRequest, LimitCreateResponse)

  limit.proto:19

  Create a new limit with a name and max value

  message LimitCreateRequest

  limit.proto:232

  Create

  • string name = 1

  • int32 max_value = 2

  • LimitType type = 3

    defaults to FRAME

  • LimitEnforcement enforcement = 4

    defaults to ENFORCED

  • int32 soft_value = 5

    0 or -1 means "same as max_value"

  • int32 exit_status = 6

    0 = no failure rule

  • int32 delay_minutes = 7

  • bool auto_tag = 8

    defaults to true when exit_status is set

  message LimitCreateResponse

  limit.proto:243

  • optional Limit limit = 1

• rpc Delete (LimitDeleteRequest, LimitDeleteResponse)

  limit.proto:22

  Delete the specified limit.

  message LimitDeleteRequest

  limit.proto:248

  Delete

  • string name = 1

  message LimitDeleteResponse

  limit.proto:252

  (message has no fields)

• rpc Find (LimitFindRequest, LimitFindResponse)

  limit.proto:25

  Find a limit by name

  message LimitFindRequest

  limit.proto:255

  Find

  • string name = 1

  message LimitFindResponse

  limit.proto:259

  • optional Limit limit = 1

• rpc Get (LimitGetRequest, LimitGetResponse)

  limit.proto:28

  Get a limit by id

  message LimitGetRequest

  limit.proto:264

  Get

  • string id = 1

  message LimitGetResponse

  limit.proto:268

  • optional Limit limit = 1

• rpc GetAll (LimitGetAllRequest, LimitGetAllResponse)

  limit.proto:31

  Get all limits

  message LimitGetAllRequest

  limit.proto:273

  GetAll

  (message has no fields)

  message LimitGetAllResponse

  limit.proto:275

  • repeated Limit limits = 1

• rpc GetBindings (LimitGetBindingsRequest, LimitGetBindingsResponse)

  limit.proto:57

  Layers bound to a limit, optionally filtered by how they were bound.

  message LimitGetBindingsRequest

  limit.proto:339

  GetBindings

  • string limit_name = 1

  • repeated LimitBindSource sources = 2

    Empty means all origins.

  • repeated string layer_ids = 3

    Empty means all layers. Filters server-side so a caller asking whether a handful of layers are bound does not pull every binding on the farm.

  message LimitGetBindingsResponse

  limit.proto:348

  • repeated LimitBinding bindings = 1

• rpc GetHolds (LimitGetHoldsRequest, LimitGetHoldsResponse)

  limit.proto:68

  Current token holders, optionally filtered to one limit.

  message LimitGetHoldsRequest

  limit.proto:380

  GetHolds

  • string limit_name = 1

    Empty means all limits.

  • string host_name = 2

    Empty means all hosts. Short name or FQDN; matched against the normalized short hostname. Filters server-side so a caller asking about one machine does not pull every holder on the farm.

  message LimitGetHoldsResponse

  limit.proto:389

  • repeated LimitHold holds = 1

• rpc Rename (LimitRenameRequest, LimitRenameResponse)

  limit.proto:34

  Rename an existing limit

  message LimitRenameRequest

  limit.proto:280

  Rename

  • string old_name = 1

  • string new_name = 2

  message LimitRenameResponse

  limit.proto:285

  (message has no fields)

• rpc ReportUsage (LimitReportUsageRequest, LimitReportUsageResponse)

  limit.proto:65

  Feed Cuebot the license server's current view. Each LimitReport fully replaces the hold set for its limit and advances its watermark.

  message LimitReportUsageRequest

  limit.proto:364

  ReportUsage

  • repeated LimitReport reports = 1

  • string source = 2

    e.g. "sesictrl@lic01"

  message LimitReportUsageResponse

  limit.proto:369

  • repeated Limit limits = 1

    Post-merge state of every limit named in the request.

  • repeated string unknown_limits = 2

    Names in the request with no matching limit. Not an error.

  • repeated LimitReportSkip skipped_limits = 3

    Limits left untouched, with the reason. The rest of the batch still applies. Not an error.

• rpc SetEnforcement (LimitSetEnforcementRequest, LimitSetEnforcementResponse)

  limit.proto:42

  Set whether the limit gates booking, only informs it, or is ignored.

  message LimitSetEnforcementRequest

  limit.proto:304

  SetEnforcement

  • string name = 1

  • LimitEnforcement enforcement = 2

  message LimitSetEnforcementResponse

  limit.proto:309

  (message has no fields)

• rpc SetFailureRule (LimitSetFailureRuleRequest, LimitSetFailureRuleResponse)

  limit.proto:54

  Set the exit status, backoff and auto-tagging behavior in one call; they are only meaningful together. exit_status = 0 clears the rule.

  message LimitSetFailureRuleRequest

  limit.proto:328

  SetFailureRule

  • string name = 1

  • int32 exit_status = 2

    0 clears the rule. Otherwise must be > 1 and unclaimed by another limit.

  • int32 delay_minutes = 3

  • bool auto_tag = 4

  message LimitSetFailureRuleResponse

  limit.proto:336

  (message has no fields)

• rpc SetMaxValue (LimitSetMaxValueRequest, LimitSetMaxValueResponse)

  limit.proto:36

  message LimitSetMaxValueRequest

  limit.proto:288

  SetMaxValue

  • string name = 1

  • int32 max_value = 2

  message LimitSetMaxValueResponse

  limit.proto:293

  (message has no fields)

• rpc SetReportTtl (LimitSetReportTtlRequest, LimitSetReportTtlResponse)

  limit.proto:50

  Set the seconds after which an external report is considered stale. 0 disables staleness entirely (internal-only limit).

  message LimitSetReportTtlRequest

  limit.proto:320

  SetReportTtl

  • string name = 1

  • int32 report_ttl = 2

  message LimitSetReportTtlResponse

  limit.proto:325

  (message has no fields)

• rpc SetSoftValue (LimitSetSoftValueRequest, LimitSetSoftValueResponse)

  limit.proto:46

  Set the threshold above which only hosts already holding a token may book. 0 or -1 means "same as max_value". HOST limits only.

  message LimitSetSoftValueRequest

  limit.proto:312

  SetSoftValue

  • string name = 1

  • int32 soft_value = 2

    0 or -1 means "same as max_value"

  message LimitSetSoftValueResponse

  limit.proto:317

  (message has no fields)

• rpc SetType (LimitSetTypeRequest, LimitSetTypeResponse)

  limit.proto:39

  Set how the limit counts tokens: per frame or per host.

  message LimitSetTypeRequest

  limit.proto:296

  SetType

  • string name = 1

  • LimitType type = 2

  message LimitSetTypeResponse

  limit.proto:301

  (message has no fields)

Used in: job.LayerGetLimitsResponse, LimitCreateResponse, LimitFindResponse, LimitGetAllResponse, LimitGetResponse, LimitReportUsageResponse

• string id = 1

• string name = 2

• int32 max_value = 3

• int32 current_running = 4

  Deprecated: alias of current_usage, kept for wire compatibility.

• LimitType type = 5

• LimitEnforcement enforcement = 6

• int32 soft_value = 7

  Threshold above which only hosts already holding a token may book. -1 means "same as max_value". HOST limits only.

• int32 current_usage = 8

  Merged usage: settled + pending. What the dispatcher gates on.

• int32 settled_usage = 9

  Reported by the license server as of last_report_time.

• int32 pending_usage = 10

  Booked since last_report_time and not yet visible to the server.

• int32 host_count = 11

  Distinct hosts holding at least one token.

• int64 last_report_time = 12

  Epoch seconds of the last accepted report; 0 if never reported.

• string report_source = 13

• int32 report_ttl = 14

  Seconds after which a report is stale; 0 disables staleness.

• bool report_stale = 15

• bool blocking_disabled = 16

  True when the limit is not currently blocking, for any reason (ADVISORY, DISABLED, or stale). Saves clients recomputing the rule.

• int32 exit_status = 17

  Frame exit status meaning "this license was unavailable". 0 = no rule. Replaces the dispatcher.layer_delay.rules property. Must be > 1.

• int32 delay_minutes = 18

  Minutes to postpone a layer whose frame reported exit_status. 0 = never.

• bool auto_tag = 19

  Bind the failing frame's layer to this limit on exit_status.

• int32 spec_layer_count = 20

  Layers currently bound to this limit, split by how they were bound.

• int32 auto_layer_count = 21

How a layer came to be bound to a limit.

Used in: LimitBinding, LimitClearBindingsRequest, LimitGetBindingsRequest

• SPEC = 0

  Declared in the job spec at launch. Never touched by automation.

• AUTO = 1

  Inferred from a frame failing with the limit's exit_status.

• MANUAL = 2

  Added by an operator.

A layer bound to a limit.

Used in: LimitGetBindingsResponse

• string limit_id = 1

• string layer_id = 2

• string layer_name = 3

• string job_name = 4

• repeated string services = 5

  Services declared on the layer, useful for spotting promotion candidates.

• LimitBindSource source = 6

• int64 create_time = 7

Whether a limit gates booking, or only informs it.

Used in: Limit, LimitCreateRequest, LimitSetEnforcementRequest

• ENFORCED = 0

  Blocks booking at the thresholds. Default; pre-existing behavior.

• ADVISORY = 1

  Never blocks. Still biases dispatch toward hosts already holding a token, and still reports usage. Use when the license server does the enforcing.

• DISABLED = 2

  Ignored by the dispatcher entirely; usage is still reported.

A host currently holding at least one token of a limit.

Used in: LimitGetHoldsResponse

• string limit_id = 1

• string limit_name = 2

• string host_name = 3

  Normalized short hostname, lowercased. The join key.

• string host_id = 4

  Cuebot host id, empty when the holder is not a registered render host.

• int32 tokens = 5

• string user = 6

• LimitHoldSource source = 7

• int64 report_time = 8

Where a token hold came from.

Used in: LimitHold

• CUE = 0

  a Cue frame is running on this host

• EXTERNAL = 1

  the license server reports this host holding

• BOTH = 2

One host's checkout count as observed by the license server.

Used in: LimitReport

• string host_name = 1

  Hostname as the license server reports it. Cuebot normalizes it.

• int32 tokens = 2

  must be >= 1; omit rows with 0

• string user = 3

A license server's full snapshot of one limit's holders.

Used in: LimitReportUsageRequest

• string limit_name = 1

• repeated LimitHostUsage hosts = 2

  Full snapshot of holders for this limit at the time of capture. An empty list is meaningful: it clears the hold set.

• int32 total_licenses = 3

  Optional. When > 0, also updates max_value so the pool size tracks the license server. When 0 or unset, max_value is left alone.

• int64 capture_time = 4

  Optional. Epoch seconds the license server view was captured. When 0, Cuebot stamps receipt time. Supplying it makes the watermark accurate when the reporter itself is slow.

A limit the report left untouched, and why.

Used in: LimitReportUsageResponse

• string limit_name = 1

• LimitReportSkipReason reason = 2

Why a limit named in a usage report was not applied. Neither is an error.

Used in: LimitReportSkip

• RATE_LIMITED = 0

  Another reporter posted inside limit.min_report_interval_seconds.

• OUT_OF_ORDER = 1

  The snapshot is older than the limit's current watermark: the reporter's clock is behind, or a queued snapshot was replayed out of order. Applying it would move the watermark backwards and read as stale immediately.

How a limit counts tokens against its maximum.

Used in: Limit, LimitCreateRequest, LimitSetTypeRequest

• FRAME = 0

  Each running frame consumes one token. Default; pre-existing behavior.

• HOST = 1

  All frames on a single host share one token. For per-machine licenses such as Houdini/sesinetd, Katana or Mari. For a HOST limit, max_value is the maximum number of distinct machines the farm may spread across.