Skip to content

chore(sandbox): remove runloop from active benchmark participants - #498

Merged
kisernl merged 2 commits into
masterfrom
devin/1791509894-remove-runloop
Oct 9, 2026
Merged

kisernl merged 2 commits into
masterfrom
devin/1791509894-remove-runloop

Conversation

@kisernl

@kisernl kisernl commented Oct 9, 2026 •

Copy link
Copy Markdown
Collaborator

Summary

Comments out runloop as an active benchmark participant, matching the pattern used for other paused providers (quilt, lightning, opencomputer):

  • benchmarks/sandbox/providers.ts: runloop entry + import commented out
  • benchmarks/scale/providers.ts: runloop entry + import commented out (100k burst opt-in list)
  • .github/workflows/sandbox-tti-benchmarks.yml, sandbox-dax-benchmarks.yml, sandbox-capabilities.yml: - runloop matrix rows commented out
  • package.json: dead bench:runloop script removed (selectParticipants would reject runloop anyway)
  • benchmarks/scale/scripts/start.ts: parseArgs now validates --provider via getProvider, so a removed or misspelled provider exits before a platform run or VMs are created

The benchmarks-daily repo builds its matrix from benchmarks/sandbox/providers.ts via scripts/build-daily-matrix.mjs, so runloop drops out of daily sandbox/TTI/DAX runs automatically once this merges — no benchmarks-daily change needed (verified locally: runloop is absent from the generated include matrix).

Left intact for easy re-enablement: the @computesdk/runloop dep, provider-vars.json vault keys, .env.example, and the DAX_RESOURCE_OPTIONS sizing entry.

Link to Devin session: https://app.devin.ai/sessions/d5b87f8d22e5454f87d3a79eb2c4c37f
Open in Devin Desktop: https://app.devin.ai/desktop/session/d5b87f8d22e5454f87d3a79eb2c4c37f?variant=devin
Requested by: @kisernl

Comment out the runloop provider entry and its CI matrix rows, matching
the pattern used for other paused providers (quilt, lightning, opencomputer).

The benchmarks-daily matrix is generated from this providers list, so
runloop drops out of daily sandbox/TTI/DAX runs once this lands.

Co-Authored-By: Noah Kiser <noah@computesdk.com>
@devin-ai-integration

Copy link
Copy Markdown
Contributor

I'll fix CI failures and address comments from users with write access. I'll skip comments containing "(aside)".

  • Disable automatic comment, CI, and merge conflict monitoring

@kisernl
kisernl marked this pull request as ready for review October 9, 2026 02:03
@open-cla

open-cla Bot commented Oct 9, 2026

Copy link
Copy Markdown

Contributor License Agreement

All contributors are covered by a CLA.

@devin-ai-integration devin-ai-integration Bot left a comment •

Copy link
Copy Markdown
Contributor

Choose a reason for hiding this comment

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

✅ Devin resolved all 2 findings on 9d71f5a

Fixed by Devin (2)

  • Paused scale run consumes empty VMs
  • Runloop benchmark command always fails

View all findings in Devin Review

Devin Review

…vider

- Remove the now-dead pnpm bench:runloop script (selectParticipants would
  reject runloop anyway); matches quilt/collimate/lelantos having no script.
- start.ts: validate --provider via getProvider in parseArgs so a removed
  or misspelled provider exits before a platform run or VMs are created.

Co-Authored-By: Noah Kiser <noah@computesdk.com>
@kisernl
kisernl merged commit 6983ce0 into master Oct 9, 2026
6 checks passed
@computesdk

computesdk Bot commented Oct 9, 2026

Copy link
Copy Markdown

Sandbox Benchmark Results

Burst

# Provider Score Median TTI P95 P99 Status
1 createos 99.2 0.08s 0.08s 0.08s 10/10
2 archil 98.5 0.10s 0.23s 0.23s 10/10
3 givemeanode 98.4 0.16s 0.17s 0.17s 10/10
4 arker 98.3 0.16s 0.19s 0.19s 10/10
5 miosa 98.0 0.20s 0.21s 0.21s 10/10
6 superserve 96.8 0.29s 0.37s 0.37s 10/10
7 declaw 96.7 0.31s 0.38s 0.38s 10/10
8 vercel 96.3 0.34s 0.42s 0.42s 10/10
9 daytona 95.6 0.35s 0.58s 0.58s 10/10
10 buddy 95.2 0.46s 0.51s 0.51s 10/10
11 beam 94.0 0.59s 0.61s 0.61s 10/10
12 cloudflare 93.0 0.55s 0.93s 0.93s 10/10
13 blaxel 92.4 0.75s 0.77s 0.77s 10/10
14 tenki 91.0 0.89s 0.90s 0.90s 10/10
15 modal 90.7 0.77s 1.18s 1.18s 10/10
16 sandbox0 89.9 0.98s 1.06s 1.06s 10/10
17 upstash 86.4 1.34s 1.39s 1.39s 10/10
18 microsandbox 84.9 1.30s 1.81s 1.81s 10/10
19 tensorlake 72.8 2.59s 2.91s 2.91s 10/10
20 cloud-run 69.4 2.66s 3.67s 3.67s 10/10
21 e2b 52.2 1.30s 30.16s 30.16s 10/10
22 codesandbox 0.0 20.46s 21.06s 21.06s 10/10
23 gravixlayer 0.0 0.00s 0.00s 0.00s 0/10
24 run-cloud 0.0 44.25s 54.68s 54.68s 7/10

View full run · SVGs available as build artifacts

@computesdk

computesdk Bot commented Oct 9, 2026

Copy link
Copy Markdown

Sandbox Dax Benchmark Results

Provider Phases Total Prepare Clone Install Typecheck Status
archil 0/7 0.00s 0.00s 0.00s 0.00s 0.00s 0/1 OK
arker 7/7 72.44s 4.78s 2.06s 18.66s 42.13s 1/1 OK
beam 7/7 47.05s 2.75s 1.92s 10.48s 24.68s 1/1 OK
blaxel 7/7 48.11s 1.36s 1.62s 11.26s 25.54s 1/1 OK
buddy -- 0.00s 0.00s 0.00s 0.00s 0.00s 0/1 OK
cloud-run 2/7 1.04s 0.36s 0.00s 0.00s 0.00s 0/1 OK
cloudflare 7/7 154.54s 2.74s 4.69s 51.63s 69.25s 1/1 OK
codesandbox 0/7 0.00s 0.00s 0.00s 0.00s 0.00s 0/1 OK
createos 7/7 50.66s 2.70s 2.18s 11.80s 25.39s 1/1 OK
daytona 7/7 74.02s 6.25s 1.84s 14.29s 36.83s 1/1 OK
declaw 7/7 206.50s 5.79s 4.83s 30.44s 151.68s 1/1 OK
e2b 7/7 81.12s 11.00s 3.00s 19.75s 38.05s 1/1 OK
freestyle 0/7 0.00s 0.00s 0.00s 0.00s 0.00s 0/1 OK
givemeanode 7/7 39.06s 1.40s 1.17s 10.99s 22.52s 1/1 OK
hopx -- 0.00s 0.00s 0.00s 0.00s 0.00s 0/1 OK
isorun 7/7 31.79s 2.49s 0.96s 8.32s 17.07s 1/1 OK
microsandbox 7/7 48.60s 2.11s 2.07s 13.16s 24.24s 1/1 OK
miosa 7/7 84.45s 2.05s 5.25s 22.73s 48.21s 1/1 OK
modal 7/7 129.58s 3.58s 3.26s 27.02s 86.44s 1/1 OK
mosaic 7/7 55.60s 0.90s 2.89s 12.62s 28.39s 1/1 OK
namespace 6/7 37.33s 5.00s 0.93s 8.78s 0.00s 0/1 OK
northflank -- 0.00s 0.00s 0.00s 0.00s 0.00s 0/1 OK
run-cloud -- 0.00s 0.00s 0.00s 0.00s 0.00s 0/1 OK
runloop 7/7 89.42s 17.75s 4.93s 13.75s 47.12s 1/1 OK
sail 7/7 72.71s 4.74s 4.75s 18.84s 33.33s 1/1 OK
sandbox0 7/7 76.01s 2.51s 1.96s 29.08s 37.95s 1/1 OK
superserve 7/7 105.94s 11.18s 4.29s 17.80s 67.92s 1/1 OK
tenki 7/7 78.57s 4.55s 3.24s 21.68s 38.25s 1/1 OK
tensorlake 7/7 57.51s 11.45s 1.44s 10.38s 29.06s 1/1 OK
upstash 7/7 55.15s 4.36s 1.85s 11.38s 26.62s 1/1 OK
vercel 7/7 71.86s 11.47s 2.02s 12.03s 38.96s 1/1 OK
opencomputer -- 0.00s 0.00s 0.00s 0.00s 0.00s 0/1 OK

View full run

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.

1 participant