Aliaksandr Valialkin

The overall design looks good, except of the following issue: it would be more consistent with [per-tenant retention](https://docs.victoriametrics.com/cluster-victoriametrics/#retention-filters) and [per-tenant downsampling](https://docs.victoriametrics.com/cluster-victoriametrics/#downsampling) to pass filters on tenants via `vm_account_id` and `vm_project_id`...

VictoriaMetrics buffers incoming data in memory and flushes it to disk every second. This is needed for improving data ingestion performance. Cluster version of VictoriaMetrics has additional in-memory buffers for...

> @valyala do you see any implications of allowing to use 0 as search.latencyOffset value and prevent it from falling back to 1s? I think it is OK to allow...

VictoriaMetrics allows setting zero value for `-search.latencyOffset` command-line flag starting from [v1.88.0](https://docs.victoriametrics.com/CHANGELOG.html). Closing the issue as resolved then.

FYI, [vmctl](https://docs.victoriametrics.com/vmctl.html) is able to migrate data from any system, which supports Prometheus remote_read protocol, to VictoriaMetrics, starting from [v1.85.0](https://github.com/VictoriaMetrics/VictoriaMetrics/releases). See [these docs](https://docs.victoriametrics.com/vmctl.html#migrating-data-by-remote-read-protocol).

This sounds like a good feature request.

> what about only recording slow queries? since record all the running queries may use more memories and has more operations Slow queries are already logged if they take longer...

FYI, the latest commits in [master](https://github.com/VictoriaMetrics/VictoriaMetrics/tree/master) and [cluster](https://github.com/VictoriaMetrics/VictoriaMetrics/tree/cluster) branches implement `/api/v1/status/active_queries` page, which exposes currently executed queries. These commits will be included in the next release of VictoriaMetrics.

The `/api/v1/status/active_queries` page is available starting from [v1.38.0](https://github.com/VictoriaMetrics/VictoriaMetrics/releases). Leaving the feature request open until the ability to stop a specific query is implemented.

Grafana and [vmui](https://docs.victoriametrics.com/#vmui) use the following APIs for auto-completion: - [`/api/v1/labels`](https://docs.victoriametrics.com/url-examples/#apiv1labels) - [`/api/v1/label/.../values`](https://docs.victoriametrics.com/url-examples/#apiv1labelvalues) - [`/api/v1/series`](https://docs.victoriametrics.com/url-examples/#apiv1series) These APIs are enough for auto-completion tasks. So closing this feature request as won't implement....