GitHub

ShardingConfigService

API reference for ShardingConfigService.

A runner's sharding settings. runner_address None makes a client-only runner; runner_listen_address is where it binds, when not at runner_address.

Attributes

AttributeTypeDescription
runner_addressRunnerAddress | None
runner_listen_addressRunnerAddress | None
runner_shard_weightint
available_shard_groupstuple[str, ...]
assigned_shard_groupstuple[str, ...]
shards_per_groupint
shard_lock_refresh_intervalDuration
shard_lock_expirationDuration
shard_lock_disable_advisorybool
preemptive_shutdownbool
entity_mailbox_capacityint | Unbounded
entity_mailbox_metricsbool
max_resident_entitiesint | Unbounded
unprocessed_message_batch_sizeint
entity_max_idle_timeDuration
entity_registration_timeoutDuration
entity_termination_timeoutDuration
entity_message_poll_intervalDuration
entity_reply_poll_intervalDuration
refresh_assignments_intervalDuration
send_retry_intervalDuration
runner_join_debounceDuration

How long the healthy runners must stay unchanged before runners that joined get shards, so a burst of joins hands each shard off once. Only joins wait, at most four times this long; 0 publishes every change at once.

runner_health_check_intervalDuration

How often runners are health checked (default 1 minute); unhealthy ones lose their shards. The checks run on one runner, the host of the runner health singleton, so they cannot see that runner crash: it keeps its shards until runner storage stops listing it, which for SQL runner storage is once its last heartbeat is older than shard_lock_expiration. To notice a crash of that host sooner, lower shard_lock_expiration, at the cost of less tolerance for slow or unavailable lock storage.

simulate_remote_serializationbool

Functions

__init__

__init__(*, runner_address, runner_listen_address, runner_shard_weight, available_shard_groups, assigned_shard_groups, shards_per_group, shard_lock_refresh_interval, shard_lock_expiration, shard_lock_disable_advisory, preemptive_shutdown, entity_mailbox_capacity, entity_mailbox_metrics, max_resident_entities, unprocessed_message_batch_size, entity_max_idle_time, entity_registration_timeout, entity_termination_timeout, entity_message_poll_interval, entity_reply_poll_interval, refresh_assignments_interval, send_retry_interval, runner_join_debounce, runner_health_check_interval, simulate_remote_serialization) -> None
ParameterTypeDescription
runner_addressRunnerAddress | None
runner_listen_addressRunnerAddress | None
runner_shard_weightint
available_shard_groupstuple[str, ...]
assigned_shard_groupstuple[str, ...]
shards_per_groupint
shard_lock_refresh_intervalDuration
shard_lock_expirationDuration
shard_lock_disable_advisorybool
preemptive_shutdownbool
entity_mailbox_capacityint | Unbounded
entity_mailbox_metricsbool
max_resident_entitiesint | Unbounded
unprocessed_message_batch_sizeint
entity_max_idle_timeDuration
entity_registration_timeoutDuration
entity_termination_timeoutDuration
entity_message_poll_intervalDuration
entity_reply_poll_intervalDuration
refresh_assignments_intervalDuration
send_retry_intervalDuration
runner_join_debounceDuration
runner_health_check_intervalDuration
simulate_remote_serializationbool