Skip to content

Pull requests: llm-d/llm-d-kv-cache

New pull request New
32 Open 445 Closed
32 Open 445 Closed
Author
Filter by author
Loading
Label
Filter by label
Loading
Use alt + click/return to exclude labels
or ⇧ + click/return for logical OR
Projects
Filter by project
Loading
Milestones
Filter by milestone
Loading
Reviews
Filter by reviews
No reviews Review required Approved review Changes requested
Assignee
Filter by who’s assigned
Assigned to nobody Loading
Sort
Sort by
Newest Oldest Most commented Least commented Recently updated Least recently updated Best match
Most reactions
👍 👎 😄 🎉 😕 ❤️ 🚀 👀

Pull requests list

Add license scan report and status size/XS Denotes a PR that changes 0-9 lines, ignoring generated files.
#696 opened Jul 8, 2026 by fossabot Loading…
1
feat(tier-normalization): normalize device tier names and default unknown tier weight to 0 size/L Denotes a PR that changes 100-499 lines, ignoring generated files.
#695 opened Jul 8, 2026 by Alex-ai-future Contributor Loading…
4
deps(go): bump the go-dependencies group across 1 directory with 5 updates dependencies Pull requests that update a dependency file size/M Denotes a PR that changes 30-99 lines, ignoring generated files.
#694 opened Jul 7, 2026 by dependabot Bot Loading…
1
fix(pvc_evictor): restart crawler processes on unexpected death size/S Denotes a PR that changes 10-29 lines, ignoring generated files.
#692 opened Jul 6, 2026 by Alex-ai-future Contributor Loading…
1 of 2 tasks
feat: replay buffer consumer for KV event index recovery size/XL Denotes a PR that changes 500-999 lines, ignoring generated files.
#691 opened Jun 30, 2026 by RishabhSaini Contributor Loading…
2 tasks done
1
1
deps(actions): bump actions/cache from 4 to 6 dependencies Pull requests that update a dependency file lgtm Looks good to me, indicates that a PR is ready to be merged. size/XS Denotes a PR that changes 0-9 lines, ignoring generated files.
#690 opened Jun 30, 2026 by dependabot Bot Loading…
2
fix(tokenizer): use writable cache paths size/M Denotes a PR that changes 30-99 lines, ignoring generated files.
#689 opened Jun 29, 2026 by xuhui-lu Loading…
1
2
fix(kvblock): stop Lookup at first prefix-chain break size/M Denotes a PR that changes 30-99 lines, ignoring generated files.
#688 opened Jun 29, 2026 by Iceber Contributor Loading…
4
refactor(tokenization): drop dead offsets return from Render hold PRs that are blocked on design, other features, release cycle, etc. size/M Denotes a PR that changes 30-99 lines, ignoring generated files.
#683 opened Jun 23, 2026 by mayur-tolexo Loading…
1
6
deps(actions): bump actions/checkout from 4 to 7 dependencies Pull requests that update a dependency file lgtm Looks good to me, indicates that a PR is ready to be merged. size/S Denotes a PR that changes 10-29 lines, ignoring generated files.
#681 opened Jun 23, 2026 by dependabot Bot Loading…
3
Trace tokenization and propagate request context size/L Denotes a PR that changes 100-499 lines, ignoring generated files.
#679 opened Jun 22, 2026 by satyamg1620 Loading…
7
fix: Add span for events process and decode size/L Denotes a PR that changes 100-499 lines, ignoring generated files.
#678 opened Jun 22, 2026 by gyliu513 Member Loading…
feat(telemetry): correlate kv-cache logs with traces via span context size/L Denotes a PR that changes 100-499 lines, ignoring generated files.
#677 opened Jun 21, 2026 by satyamg1620 • Draft
1
1
Add Prometheus alerting rules for KV-cache metrics size/XL Denotes a PR that changes 500-999 lines, ignoring generated files.
#676 opened Jun 20, 2026 by satyamg1620 Loading…
1
5
fix(kvcache): normalize storage tier scoring weights size/M Denotes a PR that changes 30-99 lines, ignoring generated files.
#675 opened Jun 17, 2026 by Iceber Contributor Loading…
4
fix(examples): wire online kv-events ZMQ subscriber lgtm Looks good to me, indicates that a PR is ready to be merged. size/M Denotes a PR that changes 30-99 lines, ignoring generated files.
#674 opened Jun 17, 2026 by Iceber Contributor Loading…
6
fix(kvblock): avoid full keyspace scan in Redis Clear size/XL Denotes a PR that changes 500-999 lines, ignoring generated files.
#673 opened Jun 16, 2026 by yankay Collaborator Loading…
1
7
fix(kvevents): support narrow integer types in getHashAsUint64 lgtm Looks good to me, indicates that a PR is ready to be merged. size/M Denotes a PR that changes 30-99 lines, ignoring generated files.
#671 opened Jun 16, 2026 by Alex-ai-future Contributor Loading…
1
7
[WIP] feat: Add storage checkpoint index for KV offloads size/XL Denotes a PR that changes 500-999 lines, ignoring generated files.
#668 opened Jun 15, 2026 by albertoperdomo2 Contributor • Draft
1
fix(kvblock): hold PodCache.mu during Lookup for atomic check-and-set lifecycle/stale size/M Denotes a PR that changes 30-99 lines, ignoring generated files.
#667 opened Jun 15, 2026 by Iceber Contributor Loading…
4
feat(telemetry): standardize OTel instrumentation scope and span names lgtm Looks good to me, indicates that a PR is ready to be merged. size/L Denotes a PR that changes 100-499 lines, ignoring generated files.
#654 opened Jun 10, 2026 by Alex-ai-future Contributor Loading…
1
19
feat(kvcache): HMA-aware KV block scoring with window-aware SWA hold PRs that are blocked on design, other features, release cycle, etc. size/XXL Denotes a PR that changes 1000+ lines, ignoring generated files.
#650 opened Jun 9, 2026 by sagearc Collaborator Loading…
12
Add subscriber and pool health metrics size/L Denotes a PR that changes 100-499 lines, ignoring generated files.
#646 opened Jun 8, 2026 by satyamg1620 Loading…
1
8
feat(kvevents): trace raw message processing lifecycle/stale size/L Denotes a PR that changes 100-499 lines, ignoring generated files.
#640 opened Jun 6, 2026 by larainema Contributor Loading…
1
1
fix(make): export EXAMPLE variable for sub-make invocations lgtm Looks good to me, indicates that a PR is ready to be merged. size/XS Denotes a PR that changes 0-9 lines, ignoring generated files.
#633 opened Jun 4, 2026 by Alex-ai-future Contributor Loading…
3
ProTip! Filter pull requests by the default branch with base:main.