This guide introduces the cluster limit designed to enhance user experience when cluster resources are limited. It sets a limit on the number of actors per parallelism (i.e., CPU core). Exceeding this limit when you create materialized view, table, or sink will trigger a warning message.
meta_actor_cnt_per_worker_parallelism_soft_limit
, defaults to 100.
If exceeding this limit, creating materialized view, table, or sink can still succeed but will generate a SQL notice message when the statement returns.meta_actor_cnt_per_worker_parallelism_hard_limit
, defaults to 400.
If exceeding this limit, creating materialized view, table, or sink will fail and generate an error message when the statement returns.If it’s safe for the cluster to operate with limits exceeded, you can override the default behavior using one of the following methods:
SET bypass_cluster_limits TO true
.Please be aware that once you bypass the check or increase the limits, the cluster could become overloaded, leading to issues with stability, availability, or performance.
Create a materialized view exceeding the soft limit
Create a materialized view exceeding the hard limit
This guide introduces the cluster limit designed to enhance user experience when cluster resources are limited. It sets a limit on the number of actors per parallelism (i.e., CPU core). Exceeding this limit when you create materialized view, table, or sink will trigger a warning message.
meta_actor_cnt_per_worker_parallelism_soft_limit
, defaults to 100.
If exceeding this limit, creating materialized view, table, or sink can still succeed but will generate a SQL notice message when the statement returns.meta_actor_cnt_per_worker_parallelism_hard_limit
, defaults to 400.
If exceeding this limit, creating materialized view, table, or sink will fail and generate an error message when the statement returns.If it’s safe for the cluster to operate with limits exceeded, you can override the default behavior using one of the following methods:
SET bypass_cluster_limits TO true
.Please be aware that once you bypass the check or increase the limits, the cluster could become overloaded, leading to issues with stability, availability, or performance.
Create a materialized view exceeding the soft limit
Create a materialized view exceeding the hard limit