Error and Log Codes

Fix the callback named by details.source

Pass the object received from Fluxer and supported size and format options

Pass attachments received from this client, and raise maxBytes for tooLarge

Read the hint, and apiError or details.providerCode for a Fluxer rejection. Repeat the call only when errors.isRetryable returns true

Check the whole bot token and the application owner's account standing. A valid token can be rejected if that account is closed or disabled, which a new token cannot fix. The rejection alone does not identify the cause. The SDK does not retry it

Read the hint, and apiError or details.providerCode for a Fluxer rejection. Repeat the call only when errors.isRetryable returns true

No action is needed during shutdown. Otherwise check what aborted the signal

Read the hint, and apiError or details.providerCode for a Fluxer rejection. Repeat the call only when errors.isRetryable returns true. Choose a compatible follow destination. Before converting a text channel to an announcement channel, remove its channel-follower webhooks

Call connect or run once per client and await that call

Create a new client to connect again. When details.outcome is unknown for a write, check whether Fluxer applied the change before repeating it. A read, marked by details.read, is safe to repeat

Start collectors after the client is ready, start a new one after a lost connection, fix a failing filter or callback, or raise the limit named in the message

Fix the named setting. The hint suggests the closest supported option for a misspelled key

Check network access to the Fluxer instance. Follow the hint for rejected tokens and shard close codes

Check network access, or raise the timeout that expired, usually connection.startupTimeoutMs

Connect the client first, and retry after a timeout or lost connection

Keep the bot's subscriptions open while it runs. The hint explains how subscriptions are numbered, or how to keep up with the events after an overflow

Handle events faster, raise the pending limits or choose a dropping overflow policy

Await the pending next call before calling next again

Raise timeoutMs, or make the filter return true or false synchronously

Send only application opcodes to a shard this client owns after it is ready

Read the hint, and apiError or details.providerCode for a Fluxer rejection. Repeat the call only when errors.isRetryable returns true

Pass a value of the documented form

Connect the client first, read batches promptly, and wait retryAfterMs after rateLimit

Check selectedMessageIds and terminalBatchIds before running the cleanup again

Read the hint, and apiError or details.providerCode for a Fluxer rejection. Repeat the call only when errors.isRetryable returns true. A send whose outcome is unknown may already have created the message

Read the hint, and apiError or details.providerCode for a Fluxer rejection. Repeat the call only when errors.isRetryable returns true. For publishing preconditions, check the source channel and message. For rate limits, wait at least retryAfterMs before repeating the call

Read the hint, and apiError or details.providerCode for a Fluxer rejection. Repeat the call only when errors.isRetryable returns true

Retry indexing later, and raise maxPages for pageLimit

Correct the input named by inputValidation.path, or select fewer members

Wait at least retryAfterMs before connecting again

Read the hint, and apiError or details.providerCode for a Fluxer rejection. Repeat the call only when errors.isRetryable returns true

Fix the callback or getter that raised the cause

Report it with the output of describeError

Read the cause and its hint

Start the child only through the supervisor

Read the supervisor log records for the child that failed

Check that the code under test sends the request, or raise timeoutMs

Fix the failing code, assert the failures before shutdown, or handle them with onError

Read the hint, and apiError or details.providerCode for a Fluxer rejection. Repeat the call only when errors.isRetryable returns true

Read the hint, and apiError or details.providerCode for a Fluxer rejection. Repeat the call only when errors.isRetryable returns true. Choose a compatible destination for a follower webhook. For a published-message edit rate limit, wait at least retryAfterMs before editing again

A cache.onChange listener had its maximum number of unfinished calls and waiting changes, so the oldest waiting change was dropped, or the client shut down while changes waited for its unfinished calls, so those were dropped. The record names the observer and how many changes it has dropped

Make the listener finish its calls, or raise its concurrency or maxPendingChanges

A cache maxAgeMs function or a cache.onChange listener threw or returned an invalid value

Fix that function or listener

A message cleanup progress callback threw

Fix the onProgress callback

A collector callback threw or failed

Fix the callback

A full collector ignored an event while its last callback finished

A collector stopped because of a failure

Read the collector's result for the error

A collector filter threw or returned a non-boolean

Make the filter return true or false without throwing

A command finished

A command threw or failed

Fix the command, or handle the failure with the router's onError

The bot user ID needed for mention prefixes could not be read, so mention commands are skipped for a while

A command did not run because middleware did not continue

A command did not run because a guard denied it, its cooldown was active or an argument was invalid

A message with the command prefix did not run a command. fields.reason holds the cause: an unknown name (the closest command is in fields.suggestion), a group without a subcommand or text the parser rejected, whose explanation the message gives

A subscription stopped with events still waiting in its queue, so they were discarded unhandled. Info means the client was closing, and Debug means the subscription itself closed. The count is in fields.discarded and in the eventsDropped.closed counter

Shut down with drainMs to let waiting handler events run first

A full subscription queue dropped an event under its overflow policy

Handle events faster or raise the pending limits

An event handler threw or failed. Other events keep running

Fix the handler, or handle the failure with onError

An onError hook failed while it handled another failure

Fix the onError hook

A subscription's queue exceeded its limit, so the subscription stopped

Handle events faster, raise the pending limits or choose a dropping overflow policy

A raw dispatch is being delivered to raw subscribers

A draining shutdown did not deliver events that open subscriptions or collectors would have received, because it accepts no new events. The count is in fields.refused and in the eventsDropped.closed counter

A subscription was registered after shutdown began, so it starts closed

Register handlers before shutting down

A command passed to gateway.send was sent to the gateway

A dispatch event was received

A dispatch failed validation and was skipped. The failing field is in fields.field

Report it if Fluxer's payloads changed, and include the field

Fluxer acknowledged a heartbeat

A shard asked Fluxer not to send dispatch types that nothing uses

A handler was registered for an event type that the running shards asked Fluxer to suppress

Register the handler before connecting, or remove the type from gateway.ignoredEvents

The member attached to a message, reaction or typing dispatch failed validation, so the event arrives without it and the member cache dropped its entry for the account. Logged only while the member cache is enabled

Report it if Fluxer's payloads changed, and include the dispatch type

The masked body of a received gateway frame, printed only in unsafe payload mode

The masked body of a sent gateway frame, printed only in unsafe payload mode

A gateway frame was received

A gateway frame was sent

A thread in a community snapshot failed validation, so the channel cache does not hold it. The failing entry is in fields.field

Report it if Fluxer's payloads changed, and include the field

A dispatch type that this SDK version does not handle was ignored

An opcode that this SDK version does not handle was ignored

A shard started a connection attempt

Automatic sharding counted the bot's communities and chose the shard count from that number

The runBot runner stopped because of a failure that no earlier record reported, and set the process exit code to 1. Setting reportFailure to false turns both off

Read the error's hint

Shards resumed sessions from the session store, which carry no community data, so the SDK refilled the enabled community, role and channel caches through REST. Warn means some communities failed or the refill stopped early, such as when the community list repeated a page

At Warn, read the message and any error. The affected cache entries fill as events and requests arrive

A cleanup step failed during shutdown. Shutdown still ran the remaining steps

Report it with the record's error if the SDK owns the failing step

Every shard of the client is connected, with the bot's name and the number of communities it is in. Warn means the bot is in no community yet

At Warn, invite the bot to a community with the installation link in the message

A shard stopped for good, for example after a rejected token, an invalid shard setting or an exhausted recovery schedule

Read the error's hint

An established shard connection was lost and the SDK is reconnecting

No action is needed unless it repeats

The Fluxer instance announced a web domain migration. When it is switched on, links and OAuth URLs use the announced web app host (Info). A switched-off migration is Debug

The drain time ran out, so shutdown cancels the handlers, waiting events, scheduled tasks and REST requests that had not finished. The fields name how many

Make handlers and tasks finish sooner, or raise drainMs if the process may take longer to stop

All work running when the drain began finished in time, so shutdown continues without cancelling any

A shutdown with drainMs stopped accepting new events and waits for running handlers, waiting handler events, running scheduled tasks and REST requests to finish

The sharding.identify coordinator failed to let a shard start a new session, so the SDK retries the connection

Fix the coordinator's permit function

A connection-state observer was still busy, so an intermediate state was replaced by a newer one

A connection-state observer threw or rejected

Fix the observer callback

A shard is ready and receives events

Fluxer closed a shard with 4011 (sharding required) on an automatically sized client or a client without sharding settings, so the SDK counted the communities again and moved every shard to a larger automatic plan. Every shard started a new session, so events sent during the move were missed

No action is needed. If it repeats, the bot may be growing fast, and each move is logged with the new total. A client without sharding settings moves again at each start while the bot has outgrown one shard, which sharding "auto" avoids

A shard's connection attempt failed and will be retried after the delay

Check network access if it repeats

The session store's load function threw, rejected or timed out, so the shard starts a new session

Fix the session store's load function

A shard's session can no longer resume, so the next connection starts a new session and missed events are lost

A shard is resuming a session saved by the session store

The session store's save function threw, rejected or timed out. During shutdown the next start then begins a new session, and at startup the shard starts a new session instead of resuming the loaded one

Fix the session store's save function

A shard's session was saved for a later Resume

A saved session was malformed, older than Fluxer's 60-second resume window, already resumed by an earlier start, or saved for another gateway or shard count, so the shard starts a new session

The client began shutting down

The client finished shutting down and released its resources

The client started connecting, with its SDK version and shard plan

Starting several shards may take longer than connection.startupTimeoutMs, because the SDK starts shard sessions one second apart

A stop signal or process signal asked the bot to stop

One run of a task started with client.schedule threw or failed. A repeating task keeps its schedule

Fix the task, or handle the failure with the client-level onError

Shutdown ended scheduled tasks that still had a run waiting or in progress, so those runs never happened or did not finish. The count is in fields.cancelled, and fields.running counts the runs cut off while in progress

Shut down with drainMs to let running tasks finish first, and keep work that must survive a restart outside the SDK

A rate-limit wait would pass the operation's timeout, so the request fails without waiting. When a 429 response caused it, the fields repeat what that response sent, named like RateLimitObservation

Raise timeoutMs, or send fewer requests on that route

A request was rate-limited and waits delayMs before retrying, at Warn for a second or more. The fields repeat what the 429 response sent, named like RateLimitObservation

Send fewer requests on that route if long waits repeat

A REST request failed with reason busy because the REST queue or the upload byte budget was full. It is logged at most once a minute, and every busy failure is counted in diagnostics().counters.restBusy

Send fewer requests at once, or raise the limit the message names

The masked body of a REST response, printed only in unsafe payload mode

The masked body of a REST request, printed only in unsafe payload mode

Fluxer rejected a request with HTTP 401 or 403, usually a token or permission problem that persists. The message names the target channel or community when the route has one. It is logged even when the application handles the Result. When a handler or command fails with this rejection within one second and no onError hook receives the failure, the handler's failure record reports it instead. A later handler failure also logs its own record, because the Warn has already been emitted. Webhook and OAuth clients log rejected webhook tokens and OAuth client credentials the same way

Follow fields.hint. A bot that expects these rejections can lower the rest category level

One REST attempt finished, with its route, status, duration and Fluxer error code

A successful response to an SDK operation did not match the expected shape. The failing field or check is in fields.field

Check test fixtures, or report it if Fluxer's responses changed

A read is retried after a transient failure

A typing indicator refresh of keepTyping failed, so refreshing stopped while the task keeps running and its result is not changed. The record is logged when the refresh fails, before the task settles, and the channel is in fields.channelId. A refresh stopped by client closure is not logged

Check the record's error for why the typing request failed. The indicator can expire before the task settles

An Effect log call inside the SDK

An internal timer callback failed

Report it with the record's error

FLUXERLY_DEBUG names categories that do not exist

Use the category names listed in fields.known

FLUXERLY_LOG_FORMAT or FLUXERLY_LOG_COLOR has a value that is not recognized, so it is ignored

Use one of the values in fields.accepted

Unsafe payload logging is on, so Trace records include gateway and REST payloads that can contain private content. Bot tokens and other secrets stay masked

Turn it off outside local debugging

The supervisor counted the bot's communities for totalShards auto and chose the shard count and child processes

A child printed a line that is not a log record. Standard error lines are Warn

A child process stopped unexpectedly and cannot be restarted, because restart is false or its restarts are used up, so the supervisor shuts down

Read the child's output records before this one, and raise restart.maxAttempts or remove restart: false if restarts should continue

A child process exited. Warn means it exited unexpectedly

The supervisor's identify coordinator rejected or threw instead of granting a child permission to start a new session, so the child retries as after a network failure

Fix the permit function of the identify.coordinator supervisor option

A child did not confirm sending Identify within 5 seconds of its permission, so the supervisor stops that child. Its restart follows the restart settings

Read the child's output records before this one

A supervised child asked its parent for a member request slot and got no answer, because the parent did not answer within one second or the child lost its message channel to it. The child then counted only its own member requests, so requests from other children could pass Fluxer's limit of 12 per account in 10 seconds, and Fluxer drops those without an answer. Logged once per child

Read the supervisor records about this child before this one

Reading a child's standard output or standard error failed, so the parent stops forwarding that stream while the child keeps running. The stream is in fields.stream

Read the record's error, and restart the supervisor if that child's output is needed

Fluxer closed a shard of a child with 4011 (sharding required) under totalShards auto. Warn means the supervisor stopped every child, counted the communities again and started a larger plan, so events sent during the move were missed. Error means it had already moved 3 times within an hour or reached 16,384 shards, so it shut down

For Error, check why the community count keeps growing, or choose a larger numeric totalShards

The supervisor restarts a child after a crash, within the restart budget

The supervisor started a child process

The supervisor could not start a child process

Check the entry path and execArgv

A child did not become ready within startupTimeoutMs, so the supervisor shuts down

Read the child's output records before this one, or raise startupTimeoutMs if startup needs longer

The supervisor force-terminates a child. Warn means the child did not exit within shutdownTimeoutMs of a stop request. Error means the child lost its message channel to the parent and kept running past shutdownTimeoutMs, so the supervisor shuts down

Make the child's cleanup finish sooner, or raise shutdownTimeoutMs

A test route's responder threw, so the request failed as a network error

Fix the responder

A test client left gateway sockets open after shutdown, and the test transport closed them

A test client request matched no test route, so it received 404

Add a route for the request