fix(stream): raise SSE event limit to 5 MiB and honour per-call maxEv…

[ditadi]

Problem

Analytics queries returning more than 1 MiB of JSON failed. The SSE event size limit (streamDefaults.maxEventSize) was 1 MiB, and oversized events are dropped — so a legitimate result set silently never reached the client.

The documented workaround (format="ARROW_STREAM", which bypasses SSE) isn't viable for every app.

There was also a second problem: StreamConfig.maxEventSize could be passed per call via executeStream({ stream: { maxEventSize } }), it type-checked, and it was then silently ignored. The value was only ever read in the StreamManager constructor, and BasePlugin constructs its manager with no arguments — so there was no way to raise the limit at all.

Changes

1. Default raised from 1 MiB to 5 MiB

At ~234 bytes per row (typical dashboard row: ids, timestamps, metrics, labels), this moves the ceiling from ~4.5k to ~22k rows. Measured peak heap is ~11 MB per stream, comfortable against the 6 GB of a Medium app. 5 MiB also matches the existing MAX_RESPONSE_BYTES in the database plugin.

Raised on both sides: streamDefaults.maxEventSize and connectSSE's maxBufferSize. These must stay in sync — a smaller client buffer turns a legal server event into "Buffer size exceeded" plus retries.

2. maxEventSize now honoured per call

Resolved once at stream creation and stored on the StreamEntry, instead of read off the shared StreamManager instance. One manager serves every stream a plugin opens, which is why the instance field made per-call overrides ineffective.

3. Size measured in UTF-8 bytes

The check used String.length (UTF-16 code units) while the error message said "bytes". Non-ASCII payloads (accents, CJK, emoji) could exceed the limit on the wire and still pass.

Scope

13 lines of production code across 4 files. bufferTTL, disconnectGraceMs, heartbeatInterval and maxActiveStreams are untouched — they have the same constructor-only limitation, but fixing them isn't needed here.

Tests

4 tests covering the per-call path, which had no coverage before (the constructor path did — that's how this survived). Verified by reintroducing each bug and confirming the tests fail.

pnpm -r typecheck clean; 112 stream tests passing.

[github-actions]

📦 Bundle size report

Compared against bundle-size-baseline.json (main).

@databricks/appkit

npm tarball (packed): 1.0 MB (+244 B) — gzipped download (dist + bin; excludes release-only docs/NOTICE).

dist	raw	gzip
JS (runtime)	1.0 MB (+223 B)	376 KB (+42 B)
Type declarations	385 KB	136 KB
Source maps	2.1 MB (+484 B)	704 KB (+184 B)
Other	11 KB	3.7 KB
Total	3.5 MB (+707 B)	1.2 MB (+226 B)

Per-entry composition (own code — deps external (as shipped))

Entry	Initial (gz)	Lazy (gz)	Total (gz)	node_modules (min)	Own code (min)
.	95 KB (+24 B)	2.5 KB	97 KB (+24 B)	external	311 KB (+58 B)
./beta	75 KB (+37 B)	458 B (+1 B)	75 KB (+38 B)	external	225 KB (+58 B)
./testing	17 KB	0 B	17 KB	external	50 KB
./tsdown	520 B	0 B	520 B	external	813 B
./type-generator	22 KB	0 B	22 KB	external	65 KB

Chunks:

Entry	Chunk	Load	Size (gz)
.	index.js	initial	91 KB
.	utils.js	initial	4.0 KB
.	remote-tunnel-manager.js	lazy	2.5 KB
./beta	beta.js	initial	59 KB
./beta	stream-manager.js	initial	5.8 KB
./beta	wide-event-emitter.js	initial	3.2 KB
./beta	databricks.js	initial	3.2 KB
./beta	configuration.js	initial	2.1 KB
./beta	service-context.js	initial	1.3 KB
./beta	client.js	initial	434 B
./beta	client-options.js	initial	220 B
./beta	supervisor-api.js	lazy	193 B
./beta	databricks.js	lazy	142 B
./beta	index.js	lazy	123 B
./testing	index.js	initial	17 KB
./tsdown	index.js	initial	520 B
./type-generator	index.js	initial	22 KB

@databricks/appkit-ui

npm tarball (packed): 350 KB (-59 B) — gzipped download (dist + bin; excludes release-only docs/NOTICE).

dist	raw	gzip
JS (runtime)	395 KB	132 KB
Type declarations	229 KB	84 KB (+1 B)
Source maps	766 KB (-121 B)	253 KB (-76 B)
CSS	16 KB	3.2 KB
Total	1.4 MB (-121 B)	473 KB (-75 B)

Per-entry composition (consumer bundle — deps bundled, peerDeps external)

Entry	Initial (gz)	Lazy (gz)	Total (gz)	node_modules (min)	Own code (min)
./js	5.3 KB	49 KB	55 KB	208 KB	14 KB
./js/beta	20 B	0 B	20 B	0 B	0 B
./react	432 KB	49 KB	481 KB	1.3 MB	177 KB
./react/beta	1.0 KB	0 B	1.0 KB	0 B	1.9 KB

Chunks:

Entry	Chunk	Load	Size (gz)
./js	index.js	initial	5.2 KB
./js	chunk	initial	120 B
./js	apache-arrow	lazy	49 KB
./js/beta	beta.js	initial	20 B
./react	index.js	initial	430 KB
./react	tslib	initial	2.1 KB
./react	apache-arrow	lazy	49 KB
./react/beta	beta.js	initial	1.0 KB

[github-actions]

🤖 AppKit PR bot

🔬 Run evals

Start an eval for this PR from the evals-monitor app: Go to Evals Monitor →

📦 Try this PR's app template

Scaffolds a new app from this PR's SDK build. Run it in any folder (requires the GitHub CLI — gh auth login — and the Databricks CLI):

gh run download 33631840535 -R databricks/appkit -n appkit-template-0.69.1-pr.5f175f9-fix-stream-config-per-call-overrides-567 -D appkit-pr-567 \
  && unzip -o "appkit-pr-567/appkit-template-0.69.1-pr.5f175f9-fix-stream-config-per-call-overrides-567.zip" -d "appkit-pr-567" \
  && databricks apps init --template "appkit-pr-567"

The template pins @databricks/appkit and @databricks/appkit-ui to tarballs built from this branch, so the scaffolded app runs against this PR's code.