feat(eve): let a provider say how much of an event it needs - #1823
Draft
chadhietala wants to merge 1 commit into
Draft
feat(eve): let a provider say how much of an event it needs#1823chadhietala wants to merge 1 commit into
chadhietala wants to merge 1 commit into
Conversation
Capture policy lived entirely inside the OpenTelemetry layer: recordInputs and recordOutputs were resolved per destination and enforced by a span processor, so a provider authored under agent/instrumentation/ saw the prompt, the response, and every tool payload with no lever to decline them — and eve built that projection whether or not anyone downstream would read it. A provider now declares capture. "metadata", the default, is structure, identity, usage, and timing; "content" adds what the conversation said. The bus computes the union once, and the AI SDK bridge skips building content when the union is false, so a directory in which nobody asked never serializes a prompt. Providers that did not ask are handed a stripped copy — built at most once per event and shared, since the original belongs to whoever did ask. The OpenTelemetry pipeline is one provider among them and declares its own capture from the destinations behind it, which is what keeps the two layers consistent: a pipeline whose destinations all declined asks for metadata, and the per-destination span processor keeps serving OTel-only concerns. Structure survives declining. action.completed still carries whether the tool returned or threw, so a provider counting failures never has to ask for content. Signed-off-by: Chad Hietala <chadhietala@gmail.com>
Contributor
|
The latest updates on your projects. Learn more about Vercel for GitHub.
|
Contributor
Bundle + Package Summary:
|
| Area | Metric | Baseline | Current | Delta |
|---|---|---|---|---|
| Package | Packed tarball | 7.71 MB | 7.71 MB | +1.1 kB |
| Package | Unpacked publish size | 28.88 MB | 28.88 MB | +4.1 kB |
| Package | Installed footprint | 72.31 MB | 72.31 MB | +4.1 kB |
| Package | Published files | 3086 | 3086 | 0 |
| Package | Installed files | 6794 | 6794 | 0 |
| Runtime | Unique function payloads | 2 | 2 | 0 |
| Runtime | Total function bytes | 17.07 MB | 17.07 MB | +352 B |
| Runtime | Public routes | 13 | 13 | 0 |
Changed function payloads vs instr/09-docs-and-composition (249dca0) (2)
| Function | Status | Baseline | Current | Delta | Route changes |
|---|---|---|---|---|---|
functions/__server.func |
changed | 8.53 MB | 8.53 MB | +176 B |
none |
functions/.well-known/workflow/v1/flow.func |
changed | 8.53 MB | 8.53 MB | +176 B |
none |
eve init install
| Metric | Baseline | Current | Delta |
|---|---|---|---|
| Installed footprint | 110.72 MB | 110.72 MB | +4.1 kB |
| Installed packages | 121 | 121 | 0 |
| dependencies | 4 | 4 | 0 |
| devDependencies | 2 | 2 | 0 |
| Dependency package bytes | 43.35 MB | 43.36 MB | +4.1 kB |
| devDependency package bytes | 5.04 MB | 5.04 MB | 0 B ➖ |
Build Metadata
- Preset:
vercel - Nitro:
nitro@3.0.260610-beta - Output directory:
apps/fixtures/weather-agent/.vercel/output - Build metadata timestamp: 2026-08-07T23:10:31.023Z
- Route aliases: 13 public, 1 internal (14 total aliases)
- Vercel routes in config: 16
- Severity legend: 🔴 dominant/large, 🟠 notable, 🟡 watch, ⚪ small
Package Drill-Down
Package Details
- Package:
eve@0.31.2 - Package directory:
packages/eve - Tarball: 7.71 MB (
eve-0.31.2.tgz) - Unpacked payload: 28.88 MB across 3086 published files
- Installed footprint: 72.31 MB across 6794 installed files
- Installed root package: 27.52 MB
- Installed dependencies: 44.78 MB
- Runtime dependencies: 2
- Peer dependencies: 5 (4 optional)
Installed footprint is measured from an isolated temporary npm install of the packed tarball.
Heavy installed dependencies
eve: 27.52 MB (38.1%)@rolldown/binding-linux-x64-gnu: 19.35 MB (26.8%)ai: 6.67 MB (9.2%)zod: 5.07 MB (7.0%)undici: 3.50 MB (4.8%)
Publish payload breakdown
Published file size
🔴 dist/src/compiled/shadcn-registry/index.js [################........] 9.68 MB 33.5%
🟠 dist/src/compiled/@photon-ai/chat-adapter-ime... [####....................] 2.42 MB 8.4%
🟠 dist/src/compiled/experimental-ai-sdk-code-mo... [###.....................] 1.51 MB 5.2%
🟡 dist/src/compiled/_chunks/workflow/undici-CUI... [#.......................] 508.8 kB 1.8%
🟡 dist/src/compiled/_chunks/node/undici-DWL_MYm... [#.......................] 502.4 kB 1.7%
🔴 Other published files [########################] 14.26 MB 49.4%
Installed footprint breakdown
Installed package size
🔴 eve [########################] 27.52 MB 38.1%
🔴 @rolldown/binding-linux-x64-gnu [#################.......] 19.35 MB 26.8%
🔴 ai [######..................] 6.67 MB 9.2%
🔴 zod [####....................] 5.07 MB 7.0%
🟠 undici [###.....................] 3.50 MB 4.8%
🟠 nitro [##......................] 2.41 MB 3.3%
🔴 Other installed packages [#######.................] 7.79 MB 10.8%
Runtime dependencies (2)
| Package | Range | Notes |
|---|---|---|
nitro |
3.0.260610-beta |
|
undici |
8.9.0 |
Peer dependencies (5)
| Package | Range | Notes |
|---|---|---|
@opentelemetry/api |
^1.0.0 |
optional peer |
ai |
catalog: |
|
braintrust |
^3.0.0 |
optional peer |
just-bash |
^3.0.0 |
optional peer |
microsandbox |
^0.5.0 |
optional peer |
eve init install drill-down
eve init install details
- Command:
eve init my-agent - Package manager:
npm - Installed footprint: 110.72 MB across 8670 installed files
- Installed packages: 121 total (115 transitive-only)
- dependencies: 4 direct packages totaling 43.36 MB
- devDependencies: 2 direct packages totaling 5.04 MB
- Other transitive package files: 62.32 MB
Installed footprint is measured from an isolated temporary eve init my-agent using the current packed eve tarball.
Heavy installed dependencies
@typescript/typescript-linux-x64: 27.95 MB (25.2%)eve: 27.52 MB (24.9%)@rolldown/binding-linux-x64-gnu: 19.35 MB (17.5%)zod: 9.02 MB (8.2%)ai: 6.67 MB (6.0%)
Installed footprint breakdown
Installed package size
🔴 @typescript/typescript-linux-x64 [########################] 27.95 MB 25.2%
🔴 eve [########################] 27.52 MB 24.9%
🔴 @rolldown/binding-linux-x64-gnu [#################.......] 19.35 MB 17.5%
🔴 zod [########................] 9.02 MB 8.2%
🔴 ai [######..................] 6.67 MB 6.0%
🟠 undici [###.....................] 3.50 MB 3.2%
🔴 Other installed packages [##############..........] 16.71 MB 15.1%
dependencies (4)
| Package | Range | Installed size | Share |
|---|---|---|---|
@vercel/connect |
0.4.3 |
141.0 kB | 0.1% |
ai |
^7.0.38 |
6.67 MB | 6.0% |
eve |
file:eve-0.31.2.tgz |
27.52 MB | 24.9% |
zod |
4.4.3 |
9.02 MB | 8.2% |
devDependencies (2)
| Package | Range | Installed size | Share |
|---|---|---|---|
@types/node |
24.x |
2.54 MB | 2.3% |
typescript |
7.0.2 |
2.50 MB | 2.3% |
Function Drill-Down
Payload Size Graph
Unique function payload size and share of total
🔴 functions/.well-known/workflow/v1/flow.func [########################] 8.53 MB 50.0%
🔴 functions/__server.func [########################] 8.53 MB 50.0%
Top Function Payloads
🟠 functions/.well-known/workflow/v1/flow.func • 1 public route • 8.53 MB
| Metric | Value |
|---|---|
| Public routes | /.well-known/workflow/v1/flow |
| Runtime | nodejs24.x |
| Handler | index.mjs |
| Payload | 8.53 MB |
| Function files | 8.53 MB across 44 files |
| Traced dependencies | 0 B |
| Signal | 🟠 Bundled file index.mjs is 2.33 MB (27.4%) |
🟠 🔎 Dependency Analysis
📦 Bundled files:
Bundled file size
🟠 index.mjs [########################] 2.33 MB 27.4%
🟠 _chunks/runtime-artifacts.mjs [################........] 1.63 MB 19.1%
🟡 _libs/undici.mjs [##########..............] 980.5 kB 11.5%
🟡 _chunks/sandbox.mjs [########................] 769.0 kB 9.0%
🟡 _libs/@ai-sdk/gateway+[...].mjs [####....................] 432.8 kB 5.1%
🟠 Other bundled files [########################] 2.38 MB 27.9%
🧾 Vercel Config
{
"handler": "index.mjs",
"launcherType": "Nodejs",
"shouldAddHelpers": false,
"supportsResponseStreaming": true,
"runtime": "nodejs24.x",
"maxDuration": "max",
"experimentalTriggers": [
{
"type": "queue/v2beta",
"topic": "__eve776561746865722d6167656e74_wkf_workflow_*",
"consumer": "default",
"retryAfterSeconds": 5,
"initialDelaySeconds": 0
}
],
"environment": {
"WORKFLOW_PRECONDITION_GUARD": "1"
}
}🟠 functions/__server.func • 12 public routes, 1 internal alias • 8.53 MB
| Metric | Value |
|---|---|
| Public routes | //eve/v1/callback/[token]/eve/v1/connections/[name]/callback/[token]/eve/v1/health/eve/v1/info/eve/v1/session/eve/v1/session/[sessionId]/eve/v1/session/[sessionId]/cancel/eve/v1/session/[sessionId]/clear/eve/v1/session/[sessionId]/compact/eve/v1/session/[sessionId]/reset/eve/v1/session/[sessionId]/stream |
| Internal aliases | /__server |
| Runtime | nodejs24.x |
| Handler | index.mjs |
| Payload | 8.53 MB |
| Function files | 8.53 MB across 44 files |
| Traced dependencies | 0 B |
| Signal | 🟠 Bundled file index.mjs is 2.33 MB (27.4%) |
🟠 🔎 Dependency Analysis
📦 Bundled files:
Bundled file size
🟠 index.mjs [########################] 2.33 MB 27.4%
🟠 _chunks/runtime-artifacts.mjs [################........] 1.63 MB 19.1%
🟡 _libs/undici.mjs [##########..............] 980.5 kB 11.5%
🟡 _chunks/sandbox.mjs [########................] 769.0 kB 9.0%
🟡 _libs/@ai-sdk/gateway+[...].mjs [####....................] 432.8 kB 5.1%
🟠 Other bundled files [########################] 2.38 MB 27.9%
🧾 Vercel Config
{
"handler": "index.mjs",
"launcherType": "Nodejs",
"shouldAddHelpers": false,
"supportsResponseStreaming": true,
"runtime": "nodejs24.x"
}Build Timing: e2e/fixtures/agent-tools-sandbox
This is an informational timing measurement inside eve build, from preflight through publication. Output-size measurement and profile writing are excluded.
Build mode: deployable Vercel build with sandbox template prewarm included.
- Build pipeline: 2.04 s -> 2.23 s (+189.4 ms) vs
instr/09-docs-and-composition (249dca0). - Timing is informational: shared GitHub runners are too variable for a hard timing budget.
Detailed phase timings vs `instr/09-docs-and-composition (249dca0)`
| Phase | Baseline | Current | Delta |
|---|---|---|---|
extension.check |
10.6 ms | 1.1 ms | -9.5 ms |
project.resolve |
1.3 ms | 0.6 ms | -0.7 ms |
workspace.create |
1.3 ms | 0.6 ms | -0.7 ms |
host.prepare |
241.3 ms | 206.4 ms | -34.9 ms |
vercel.service-prefix.resolve |
1.9 ms | 1.9 ms | 0.0 ms |
nitro.create |
196.8 ms | 229.1 ms | +32.3 ms |
sandbox.prewarm |
221.5 ms | 450.3 ms | +228.8 ms |
nitro.cache.prepare |
0.3 ms | 0.3 ms | 0.0 ms |
nitro.prepare |
0.7 ms | 0.7 ms | 0.0 ms |
nitro.public-assets |
0.8 ms | 0.8 ms | 0.0 ms |
nitro.prerender |
0.5 ms | 0.5 ms | 0.0 ms |
nitro.bundle |
1.34 s | 1.31 s | -25.8 ms |
nitro.cache.write |
0.3 ms | 0.3 ms | 0.0 ms |
vercel.workflow-function.materialize |
21.4 ms | 21.3 ms | -0.1 ms |
agent-summary.emit |
0.6 ms | 0.6 ms | 0.0 ms |
nitro.close |
0.2 ms | 0.2 ms | 0.0 ms |
output.publish |
3.2 ms | 3.4 ms | +0.2 ms |
workspace.remove |
2.0 ms | 1.8 ms | -0.2 ms |
This file contains hidden or bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
Add this suggestion to a batch that can be applied as a single commit.This suggestion is invalid because no changes were made to the code.Suggestions cannot be applied while the pull request is closed.Suggestions cannot be applied while viewing a subset of changes.Only one suggestion per line can be applied in a batch.Add this suggestion to a batch that can be applied as a single commit.Applying suggestions on deleted lines is not supported.You must change the existing code in this line in order to create a valid suggestion.Outdated suggestions cannot be applied.This suggestion has been applied or marked resolved.Suggestions cannot be applied from pending reviews.Suggestions cannot be applied on multi-line comments.Suggestions cannot be applied while the pull request is queued to merge.Suggestion cannot be applied right now. Please check back later.
Tenth and last PR of the instrumentation-providers stack. Base is
instr/09-docs-and-composition(#1795). Everything here is behindexperimental.instrumentationProviders, which stays off.Problem
Capture policy lived entirely inside the OpenTelemetry layer.
recordInputsandrecordOutputswere resolved per destination, unioned into the pipelinesettings, and enforced by
contentFilteringProcessor— a span processor. Twoconsequences, both of which the design doc calls out:
agent/instrumentation/that is not an OTeldestination — an audit log, a metrics counter, a hosted eval SDK with its own
transport — received the prompt, the response, and every tool argument and
result, with no lever to decline them. The span-processor lever is reachable
only from inside the OTel pipeline, so it could not serve those providers at
all.
downstream would read it. An agent tracing to a metadata-only destination
still paid to serialize each prompt.
Solution
A provider declares
capture."metadata"— the default — is structure,identity, usage, and timing.
"content"adds what the conversation said.The bus computes the union across registered providers once and exposes it as
hooks.capturesContent. The AI SDK bridge reads that flag and skipsconstructing content at all when it is false, which is what makes the saving
real rather than a filter applied after the fact. When some provider did ask,
the ones that did not are handed a stripped copy of the event — built at most
once per event and shared between them, since the original belongs to whoever
asked and must not be edited underneath them.
The OpenTelemetry pipeline is itself one provider in that list, and now declares
its own
capturefrom the destinations behind it. That is what keeps the twolayers consistent: a pipeline whose destinations all declined content asks for
"metadata"like anything else, and the per-destination span processor keepsserving OTel-only concerns, which the design explicitly preserves.
Behavior changes
inputonmodel.call.startedandaction.started,contentonmodel.call.completed, and the payloads insideaction.completed'soutput.InstrumentationActionOutput's discriminator survives declining. Only theinner
output/erroris stripped, so a provider counting tool failures neverhas to ask for content to learn that one failed. Same for
model.call.completed'sfinishReasonandusage.agent/instrumentation.tslayout is unaffected. Its only event isstep.started, which is theruntimeContextsurface rather than a bussubscriber, and the OTel provider derives its capture from the same
recordInputs/recordOutputsit already honored.Validation
pnpm fmt,pnpm lint,pnpm typecheck,pnpm guard:invariants,pnpm docs:check,pnpm test:unit(6126 passed),pnpm test:integration(617passed).
New unit coverage: the projection is absent end-to-end when no provider asked;
a metadata provider sharing a bus with a content one sees structure but not
content; and the
capturesContentunion itself.