Olly AI agent

The Ask Olly button from the trace drilldown header is now available wherever you open a span — from the APM Service Catalog drilldown and from a RUM session, page view, or error. Same trace context, same 5 suggested queries, same Cmd+I / Ctrl+I shortcut. Learn more

Admins can now monitor team-wide Olly AI consumption in the current billing cycle and cap individual users with a monthly limit — preventing any single user from depleting the team's shared quota. Open it from Settings, then Olly AI, then Usage Management. Learn more

A new Ask Olly button in the trace drilldown header opens the Olly side drawer with the current trace already in context — trace ID, selected span, service, operation, duration, and timing. Five trace-scoped suggested queries help you go straight to root cause, latency, downstream failure, deployment correlation, or a next debug step in a single click. Learn more

Olly's investigation is now visible live in the chat as an Execution plan, with each step labeled, timed independently, and expandable to reveal Olly's reasoning, tool details, and any charts or tables it produced along the way. When Olly finishes, the header switches to Execution summary and the steps stay open for review. Learn more

Olly now runs a single chat mode in the Coralogix UI — no more picker for Pro, Focus, or Fast. The single mode is built on the same skill-based architecture, defaults to GPT-5.4, and handles both quick lookups and deeper investigations. GitHub repository context is available across all chats.

The Olly API keeps focus and fast as accepted interaction_mode values for backwards compatibility — both route to skill. Migrate existing integrations when convenient. Learn more

Drive Olly from your own backend with the public Olly API. Trigger investigations, send prompts, and retrieve the query results Olly generates — all using the same models, skills, and permissions as the in-product chat. Personal and team API keys are both supported. Learn more

Admins can now define team-wide skills that Olly applies for everyone in the team — runbooks, response templates, or any reusable instructions you'd otherwise re-create as a personal skill. Manage them from the new Team tab on the Skills settings page, alongside your personal skills. Up to 10 team skills per team. Learn more

File attachments for Olly chat now accept .md files alongside the existing text-based formats (.pdf, .txt, .json, .csv). Learn more

Create your own custom skills — short Markdown files that teach Olly your team's workflows, terminology, or response style. Olly applies them when relevant, and you can also pick a skill directly in chat. Write skills from scratch in the settings UI or upload existing .md files. Up to 10 custom skills per user. Learn more

Admins can now define team-wide rules that Olly applies across every user in the team — for example, prioritizing production data or flagging security issues first. Team rules sit alongside personal user rules and take precedence when guidance conflicts. Learn more

Olly Pro mode now supports Gemini 3.1 Pro and Gemini 3 Flash. Select Gemini from the model dropdown in the Olly chat input bar while in Pro mode. See Model selection and pricing for per-model rates.

Pro mode is a new chat mode built on a skill-based architecture that applies specialized skills to each request — the same pattern modern AI agents use for complex tasks. It delivers faster and higher-quality answers than Focus mode for most observability work. Pro mode currently shares Focus mode's model list, with additional model support rolling out. Learn more

Attach files to a chat message to give Olly additional context for your prompt — for example, a screenshot of an error, a configuration file, or a CSV of metrics. Up to 5 files per message, 5 MB per file. Supports images (.jpg, .png, .svg, .web) and text-based files (.pdf, .txt, .json, .csv). Learn more

Connect a GitHub repository to Olly for code-aware investigations and root cause analysis. In Focus mode, Olly can trace logs, errors, and alerts to specific files and lines, explain what changed, and suggest targeted fixes. Learn more

New models supported: GPT-5.4 mini and Claude Sonnet 4.6. See pricing for per-model rates.

You can now pick a specific model in each mode instead of choosing a provider. Focus mode offers four higher-tier models and Fast mode offers two lower-tier models, with a recommended default in each. Your selection persists across chats and sessions. Learn more

New models supported: GPT-5.4 and GPT-5.2 in Focus mode. See pricing for per-model rates.

Olly documentation is now part of the Coralogix documentation portal. All Olly docs have been updated to reflect Olly as a built-in Coralogix feature, including new sections for Olly chat, model selection, and integrations with GitHub and Slack. Learn more

Olly is now a supported entity type in Quota Rules. Allocate units to Olly independently, set overflow behavior, and monitor enforcement from Data Flow, then Quota Rules. Learn more

Olly usage is now visible in the Data Usage page under a dedicated Olly tab. Monitor daily CU consumption, view per-user breakdowns, and track usage trends over time from Settings, then Data Usage. Learn more

New admin tools for managing your Olly organization: Usage Management, Organization Management, and Members Management. Learn more

New FAQ pages covering general questions, data sources, features, and admin tools. Learn more

New GitHub docs: GitHub Integration and GitHub Data Sources. Learn more

Updated Data Sources doc — explains how Olly selects data sources. Learn more

User Rules, Fast and Focus Modes, and OAuth Support for Olly. Learn more

Olly Autonomous AI Agent October 2025 release notes. Learn more