Skip to main content

Latest Release Note

· 7 min read
note

This release note includes features:

  • that are already integrated in latest version,
  • that will be integrated in future release.

Upgrade risk

info
  • No compatibility issue 👍
  • No breaking changes 👍

Key changes

How to upgrade

  1. Use Helm chart 6.2.0 from repository
  2. Adjust global.version field value to latest in your values.yaml
  3. Deploy
tip

See Reference documentation for details.

Versions

Spider

New versions of Spider components:

ComponentVersionDocker tag
Helm chart6.2.0-
Analysis UI15.0latest
Controllers3.0latest
Gossipers7.7latest
Gociphers1.8latest
Back office-latest
Login UI-latest
Monitoring UI-latest

Dependencies

This operator is required on the Kube where Spider is installed.

DependencyVersion
Elastic Cloud for Kubernetes2 or 3

These components are set up in the correct versions by the Helm chart:

DependencyVersionDocker tag
Elastic stack9.1*9.1
Redis8.28-alpine
Traefik3.5v3.5

(*) Version 7 is still available and compatible for clusters still using ECK 2.

Compatibility

Spider has been successfully tested under these versions of dependencies:

3rd party softwareVersion
Helm3.14+
Kube1.24+

List of changes

Helm chart

✨ New features
  • New Pg Upload service to allow uploading Pg Coms.
  • Allow deploying a different Whisperer version than Controller one
  • Optional namespace-scoped ECK operator (v3.2): set elasticsearch.eck.enabled: true to let Spider install its own ECK when none is present, or when the existing one does not watch Spider's namespace. Preflight guards abort on conflict.
⚙️ Improvements
🐞 Bug fixes
  • Fix overloading the version of Controller or Gocipher

Analysis UI

✨ New features
  • UI redesign to enhance usability
  • Multi-protocol communications view (label: "All") — unified grid, sequence diagram, waterfall and stats showing HTTP and PostgreSQL communications together.
  • Waterfall view.
  • Drag and drop of widgets in Dashboard configuration
  • Drag and drop of columns in Columns selection
  • Drag and drop of actors in Sequence Diagram
  • Color scale to numeric columns in Grid
  • Save filters associated to view when switching views
  • Option to show Template in sequence diagram
  • Horizontal scroll to sequence diagram
  • Enhance pcap filter editor in Create Whisperer & Whisperer Capture Config tabs
    • Syntax highlighting
    • Parser
    • Templates
  • Add Playground in Whisperer Config to test Templates and Tags configuration
⚙️ Improvements
  • PSQL view
    • PSQL stats in Host details tab
    • Limit status code errors to code 400, that is the only one generated by the parser
    • Add JSON download to PgComs
  • Whisperer Config Playground now validates Templates and Tags against the actual Go parsing engine (calls web-parser / pg-parser /v1/playground) instead of matching with the browser's JS regex engine — no more false positives, and regexes the parser cannot use (backreferences / lookaround) are now flagged.
  • Major frontend modernization (v15.0): React 18, MUI v7, Redux Toolkit, Luxon, Rspack
🐞 Bug fixes

Controllers

✨ New features
⚙️ Improvements
  • Rewrite in Go for better scalability and stability of integration with Kubernetes
🐞 Bug fixes

Gossipers (Whisperers)

✨ New features
⚙️ Improvements
🐞 Bug fixes

Gociphers

✨ New features
⚙️ Improvements
🐞 Bug fixes

Back office

✨ New features
  • New Pg Upload service to allow uploading Pg Coms.
  • New MultiRead service (multi-read) to search HTTP and PostgreSQL communications in a unified view. Queries a single ES alias (spider-search-multicoms) that unions the HTTP and PG communication indices.
  • Playground validation API on the Go parsers — POST /web-parser/v1/playground and POST /pg-parser/v1/playground run draft tag & template rules through the real parsing engine, returning extracted values per com and the list of rules the parser cannot use (RE2 rejects backreferences / lookaround). Backs the Whisperer Config Playground.
⚙️ Improvements
  • Allow service accounts to connect many times at once. Useful for automated tests, and surely for other stuff. Still logs a warning since last connection date could not be saved (OCC on Elastic).
  • Reduce CPU usage of status aggregation services (Whisps-Status-Agg, Ciphers-Status-Agg, Hosts-Agg). The instance sweep is now computed once per agent per polling batch instead of once per status message, reducing cost from O(N×K) to O(N+K). Whisperer batches in Hosts-Agg are also processed in parallel.
🐞 Bug fixes
  • TlsKeys-Linker no longer counts "no TCP session found for a captured TLS key" as a parsing error. It is an expected case (session not captured, expired, dropped packet…), so it stopped inflating the error counter in monitoring.

Login UI

✨ New features
⚙️ Improvements
🐞 Bug fixes

Monitoring UI

✨ New features
⚙️ Improvements
🐞 Bug fixes

Tools

✨ New features
  • spider CLI: search, stats, outliers, aggs now support network-logical and network-physical views (per-minute cluster-wide traffic between actors, scoped by --controller).
  • spider CLI: search attachments and show sidecars — primitives to check whether a workload already has a whisperer (ephemeral attachment or sidecar) before attaching a new one.
  • spider plugin: skills + agents now start broad investigations from the network-usage views and check existing coverage before proposing a fresh attach.
  • spider-mon CLI: services replicas, services load, services service_calls graphs added; api_latency_ms replaced by api_calls (latency, load, max90, errors per service|api).
  • spider-mon CLI: --service flag on services, elasticsearch, redis pages — filter every per-service graph (and the source side of calls graphs) by service name, with trailing-wildcard support (e.g. --service "Pg-*").
  • spider-mon Claude Code plugin: skills + agents to review self-monitoring, compare periods, decode alerts, propose tuning.
⚙️ Improvements
  • spider-mon CLI: elasticsearch latency and redis latency replaced by calls (latency, load, max90, errors per source|target|command).
  • spider Claude Code plugin: commands prefixed with spider- for cross-plugin consistency.
🐞 Bug fixes
  • spider-mon CLI: services CPU now sums across replicas (was avg) — matches GUI semantics. Existing CPU baselines diverge on first compare after upgrade.
  • spider-mon CLI: services load now sums real per-service inbound calls from the api index; previously bucketed on CB group field, which mixed datastore identifiers (elasticsearch, spider-*) with services.

Online documentation

Main updated parts:

API impacts

note

This section informs about any impact on Spider API, so you may adjust your scripts.

Data impacts

note

The table below tells if there are data mapping changes in Elasticsearch indices, associated or not with migrations (Yes ✅ / No ❌).
Migration are automated at upgrade time, but they may leave unattended indices that you have to remove manually.

IndexDescriptionMigration
httpcoms & pgcomsNew alias for multi-com-