Let a connection choose how edge connections are discovered by kmcginnes · Pull Request #2282 · aws/graph-explorer

kmcginnes · 2026-09-25T22:18:01Z

Description

Gremlin connections get an Edge Connection Discovery setting (Automatic, Complete, or Sampled) in the Advanced options, with what each choice costs spelled out in the dialog.
It can also be set with EDGE_CONNECTION_DISCOVERY for containers and notebooks. Unknown values are ignored with a console warning.
A forced complete is never bounded or silently sampled. If it fails, the recovery text tells the user what to change:
- Fetch timeout: "Switch Edge Connection Discovery to Automatic or Sampled in this connection's advanced options, because Automatic samples a graph this large instead of scanning it. Or raise the Fetch Timeout there, or clear it, since a complete scan may simply need longer than that allows."
- Database limit: "Switch Edge Connection Discovery to Automatic or Sampled in this connection's advanced options, because Automatic samples a graph this large instead of scanning it."
Changing the setting throws away the stored edge connections and the failure flag, since they seed the query and would otherwise make the change look like it did nothing.
The value is validated with Zod wherever it enters (the connection file and the configuration), because an unrecognized value used to pick the one combination that reproduces the original failure.

Validation

Tests for the radio group, the env var, Zod parsing of known and unknown values, discarding connections on change, and forced-complete planning and recovery text.
The top of this stack is the same tree as Ask the database for distinct edge connections instead of scanning edges #2244 merged onto current main.
pnpm checks and pnpm test clean on this layer.

Related Issues

Check List

I confirm that my contribution is made under the terms of the Apache 2.0 license.
I have verified pnpm checks passes with no errors.
I have verified pnpm test passes with no failures.
I have covered new added functionality with unit tests if necessary.
I have updated documentation if necessary.

## Description - `mapWithConcurrency` rejected on the first failure, but its other lanes kept looping and sending requests nobody was waiting for. The first rejection now stops the pool from starting new work. Callbacks already running aren't cancelled. - This also affects openCypher and SPARQL schema and edge connection fetches, which used to drain the whole queue after a failure. - Later in the stack, edge connection discovery relies on this, so abandoning an attempt actually stops it. ## Validation - New test: after one callback rejects, no further callbacks start. - `pnpm checks` and `pnpm test` clean on this layer. ## Related Issues - Part of #2141 (split out of #2244) - Layer 1 of 6. Stack, bottom first: #2281 → #2277 → #2278 → #2279 → #2280 → #2282 ### Check List - [x] I confirm that my contribution is made under the terms of the Apache 2.0 license. - [x] I have verified `pnpm checks` passes with no errors. - [x] I have verified `pnpm test` passes with no failures. - [x] I have covered new added functionality with unit tests if necessary. - [x] I have updated documentation if necessary.

## Description - Fetch Timeout and Neighbor Expansion Limit move into a collapsible "Advanced options" section of the connection dialog, so the common fields aren't buried. - The section opens by itself when the connection already overrides one of them, so an existing override is never hidden. - The trigger is a real button, so it works from the keyboard and announces whether it's expanded. - The Fetch Timeout error message and the troubleshooting guide now point at the connection's advanced options. ## Validation - Tests that the section starts collapsed, expands on click, and opens itself for a connection with a Fetch Timeout override. - `pnpm checks` and `pnpm test` clean on this layer. ## Screenshots A new connection starts with the advanced options collapsed. <img width="720" alt="New connection dialog with Advanced options collapsed" src="https://github.com/user-attachments/assets/134122d6-9eaf-40ef-ba2d-09ae22516c85" /> Expanding them shows Fetch Timeout and Neighbor Expansion Limit. <img width="720" alt="Advanced options expanded, showing Enable Fetch Timeout and Override Default Neighbor Expansion Limit" src="https://github.com/user-attachments/assets/77f7d6d5-3dce-4c6c-b33e-55eecc8b71df" /> Editing a connection that already sets a Fetch Timeout opens the section by itself. <img width="720" alt="Editing a connection with a 30000 ms fetch timeout, with Advanced options already open" src="https://github.com/user-attachments/assets/dd71d9cd-ca86-45c7-8ffd-ff0af44b387d" /> ## Related Issues - Part of #2141 (split out of #2244) - Layer 1 of 5. Stack, bottom first: #2278 → #2277 → #2279 → #2280 → #2282 ### Check List - [x] I confirm that my contribution is made under the terms of the Apache 2.0 license. - [x] I have verified `pnpm checks` passes with no errors. - [x] I have verified `pnpm test` passes with no failures. - [x] I have covered new added functionality with unit tests if necessary. - [x] I have updated documentation if necessary.

## Description This is the fix for #2141 on its own, with the existing batching and no new setting or strategy. - The old Gremlin query grouped every edge of a batch before sampling, and Neptune runs that grouping outside its engine, one item per edge. A graph with three edge types over 19.9M edges ran out of memory in about 30 seconds. - Each edge type now gets its own limited `union()` branch, so the 10,000-edge limit applies before anything is read. One limit after `hasLabel(A, B, ...)` would be shared, and a dominant type would fill it and leave the rest empty. - Results are grouped by edge type before counting, because DFE couldn't count one projection across several full branches (two took 54s, five timed out; grouping first handled ten in 9s). The anchor is `V().limit(1)` because `inject()` isn't native on Neptune, and mid-traversal `V()` because `E()` there needs TinkerPop 3.7. - Batches drop from 100 edge types to 10, so a request reads at most 100,000 edges. 100 types in one request took 116s and left a db.t3.medium refusing even single-type samples for two minutes. Other users of `DEFAULT_BATCH_REQUEST_SIZE` are unchanged. - Endpoint labels are folded and every entry is split on `::`, so a multi-label vertex keeps every label on Neptune 1.4 (one composite) and 1.3.5 (one entry per label). Before, 1.3.5 silently kept only the first. - Dedup uses `createEdgeConnectionId`, so a label containing the old `-` separator can't collide. ## Validation - The query shape matched the true edge connections on Neptune 1.2.1.0 (15/15), 1.3.5.0 (522/522), 1.4.5.1 (2,033/2,033) and 1.4.7.0 (10,018/10,018), and ran on TinkerPop 3.6.2. - Four concurrent requests of 10 full edge types each finished in 36s on db.t3.medium instances with and without DFE, and left them answering normally. - Tests: template shape and escaping, batches of 10, multi-label composites and one-per-entry labels. - `pnpm checks` and `pnpm test` clean on this layer. ## Related Issues - Fixes #2141 - Split out of #2244 - Layer 1 of 4. Stack, bottom first: #2279 → #2277 → #2280 → #2282 ### Check List - [x] I confirm that my contribution is made under the terms of the Apache 2.0 license. - [x] I have verified `pnpm checks` passes with no errors. - [x] I have verified `pnpm test` passes with no failures. - [x] I have covered new added functionality with unit tests if necessary. - [x] I have updated documentation if necessary.

kmcginnes added this pull request to stack #2283 September 25, 2026 22:25

kmcginnes force-pushed the edge-stack/6-edge-connection-discovery-setting branch from c83c5eb to ee63ec1 Compare September 25, 2026 22:46

kmcginnes force-pushed the edge-stack/6-edge-connection-discovery-setting branch 2 times, most recently from 858c1b4 to a9ce00b Compare September 25, 2026 23:22

kmcginnes force-pushed the edge-stack/6-edge-connection-discovery-setting branch from a9ce00b to d328a59 Compare September 25, 2026 23:40

kmcginnes force-pushed the edge-stack/6-edge-connection-discovery-setting branch from d328a59 to 2af664b Compare September 26, 2026 17:48

kmcginnes force-pushed the edge-stack/6-edge-connection-discovery-setting branch from 2af664b to 32c35be Compare September 26, 2026 21:57

kmcginnes removed this pull request from stack #2283 September 26, 2026 23:14

kmcginnes force-pushed the edge-stack/6-edge-connection-discovery-setting branch from 32c35be to e2829b9 Compare September 26, 2026 23:15

kmcginnes added this pull request to stack #2289 September 26, 2026 23:15

kmcginnes added 2 commits September 26, 2026 18:32

Let a connection choose how edge connections are discovered

2d0fda7

Format the edge connection discovery table for the new Markdown style

f8428f2

kmcginnes removed this pull request from stack #2289 September 26, 2026 23:34

kmcginnes force-pushed the edge-stack/6-edge-connection-discovery-setting branch from e2829b9 to f8428f2 Compare September 26, 2026 23:34

kmcginnes added this pull request to stack #2290 September 26, 2026 23:34

Navigation Menu

Let a connection choose how edge connections are discovered - #2282

Let a connection choose how edge connections are discovered#2282
kmcginnes wants to merge 2 commits into
edge-stack/5-edge-connection-discovery-strategyfrom
edge-stack/6-edge-connection-discovery-setting

kmcginnes commented Sep 25, 2026 •

edited

Loading

Uh oh!

This branch has not been deployed

Reviewers

Assignees

Labels

Projects

Milestone

Development

Uh oh!

1 participant

Conversation

kmcginnes commented Sep 25, 2026 • edited Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

Description

Validation

Related Issues

Check List

Uh oh!

This branch has not been deployed

Reviewers

Assignees

Labels

Projects

Milestone

Development

Uh oh!

1 participant

kmcginnes commented Sep 25, 2026 •

edited

Loading