Skip to content

docs: describe cluster lifecycle states on the clusters concept page - #38629

Open
maheshwarip wants to merge 3 commits into
mainfrom
claude/slack-session-x4qyu4
Open

docs: describe cluster lifecycle states on the clusters concept page#38629
maheshwarip wants to merge 3 commits into
mainfrom
claude/slack-session-x4qyu4

Conversation

@maheshwarip

Copy link
Copy Markdown
Contributor

Motivation

It is not obvious from the docs that a cluster moves through distinct lifecycle states, nor how to tell whether it is making progress (vs. stuck) in each one. This came out of a Slack discussion on cluster lifecycle and freshness/"total lag":

https://materializeinc.slack.com/archives/CU7ELJ6E9/p1788365831558719?thread_ts=1788365056.126659&cid=CU7ELJ6E9

Description

Adds a Lifecycle of a cluster section to the clusters concept page. It briefly describes the states a cluster's replicas move through and, for each state, recommends how to monitor progress:

  • Provisioning — replicas are scheduled and brought online. Monitor via mz_cluster_replica_statuses and mz_clusters.
  • Hydrating — replicas reconstruct in-memory state from Materialize's storage. Monitor via mz_hydration_statuses (and mz_compute_hydration_statuses for indexes/materialized views).
  • Catching up — the cluster works through the accumulated backlog and total lag comes down. Monitor via mz_wallclock_global_lag_recent_history and mz_materialization_lag.
  • Steady state — lag holds low and roughly constant. Monitor via the same lag relation to catch regressions.

A note also points to the snapshotting step sources go through on first start.

Verification

Docs-only change. Verified all system-catalog and concept-page links resolve against the repo, and ran bin/format-docs (no changes).

🤖 Generated with Claude Code

https://claude.ai/code/session_015919agXAtbTfF1iHNpGQoj


Generated by Claude Code

Add a "Lifecycle of a cluster" section that names the states a cluster's
replicas move through (provisioning, hydrating, catching up, steady state)
and, for each, recommends the system-catalog relation to monitor progress.

Co-Authored-By: Claude Opus 4.8 <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_015919agXAtbTfF1iHNpGQoj
@maheshwarip
maheshwarip requested a review from a team as a code owner September 2, 2026 16:23
Add documentation-authoring guidance to run doc SQL against the Materialize
emulator before committing, and to include both the query and its verified
output in syntax demos.

Co-Authored-By: Claude Opus 4.8 <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_015919agXAtbTfF1iHNpGQoj
…ates

Expand the "Lifecycle of a cluster" section with an example monitoring query
and its sample output for each state (provisioning, hydrating, catching up,
steady state). All output was captured from a running Materialize emulator.

Co-Authored-By: Claude Opus 4.8 <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_015919agXAtbTfF1iHNpGQoj
@maheshwarip
maheshwarip requested a review from antiguru September 2, 2026 20:13
@maheshwarip

Copy link
Copy Markdown
Contributor Author

Hey @antiguru would love your eyes on this PR - lmk if you disagree with this framework! Trying to put together something quick & lightweight to help users understand the lifecycle of a cluster

@antiguru antiguru left a comment

Copy link
Copy Markdown
Member

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

I think this is fine, but might need future revisions as we learn more. Let's notd block on what might be true in the future!

Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment

Labels

None yet

Projects

None yet

Development

Successfully merging this pull request may close these issues.

3 participants