Igal Shilman
Igal Shilman
Currently state mutations are always enqueued last behind any pending invocations. Operations like state mutations and possibly others, usually make very little sense to enqueue them last, but rather they...
Not a blocker, since the server will terminate forcefully, but would be nice if discovery would consider rather the server is in shutting down state or not, during retries.
Right now we excise a back off strategy per key, while it might be more beneficial also to consider a local view (per node) that keeps track of errors per...
We should make the queries to our partitioned tables (`inobx`, `outbox`, `journal`, and `state`) more efficient, if the query provides a constraint on a `service` and a `key`
The CLI currently validates that you can only `edit` state for existing-`keyed` services. But since this API calls exists on the admin api, lets make sure that the server validates...
The latest version is 34, which offers performance improvements, and bug fixes.
When setting query_parallelism = 1 we get a failure during execution with the following query: ``` RESTATE_ADMIN__QUERY_ENGINE__QUERY_PARALLELISM=1 \ RESTATE_WORKER__BOOTSTRAP_NUM_PARTITIONS=4 \ ./target/debug/restate-server --wipe all ``` ``` ./target/debug/restate sql 'SELECT ss.id FROM...
It is sometimes useful to attach additional information to a TerminalError, in addition to a textual description. especially for errors that are intended to terminate an ingress connection. Let’s consider...
Currently in the admin API, if we ask to register a deployment to an already existing lambda ARN we bump the deployment revision. We should actually fail if the lambda...