[Rust-VMM] Reduce OPS Load on Maintaining Repos
fandree at amazon.com
Tue Sep 7 09:53:44 UTC 2021
We discussed at the sync meeting about the high operational load that we currently have, and the main root cause.
One of the pain points is that we have many repositories we need to maintain, which need regular CI updates. There is also an associated cost with merging dependabot PRs. We want to improve the load by working on 3 areas:
- [DONE] archive unused repositories; repositories that do not yet have useful code are now archived; repositories to which this applies: io-rate-limiter, vmm-vcpu, vm-allocator, and kvm; once we start working on these crates, we will undo the archive
- [NOT STARTED] reduce the number of active repositories by grouping together crates in workspaces (crates will still be published independently, but they'll share the CI); this is currently pending further analysis, and we we'll come back to it in the near future.
- [IN PROGRESS] reduce the frequency of dependabot PRs from daily to weekly; this needs configuration updates in all existing repositories. I created a dummy bot that submits PRs with the changes, but there is a bug in the bot that I'll need to address before we can merge them.
Grouping repositories and archiving unused repositories helps because with a reduced number of repositories we have less dependabot PRs that we need to review. Also, when changes are required in Buildkite, we have less manual configurations to do.
What other improvements do you think would be worth to consider?
Amazon Development Center (Romania) S.R.L. registered office: 27A Sf. Lazar Street, UBC5, floor 2, Iasi, Iasi County, 700045, Romania. Registered in Romania. Registration number J22/2621/2005.
-------------- next part --------------
An HTML attachment was scrubbed...
More information about the Rust-vmm