Troubleshooting Agent overview
The Troubleshooting Agent investigates technical issues across your code, logs, databases and tickets, and explains the likely cause with its evidence.
Some tickets cannot be answered from knowledge. Something is actually broken, and finding out what means looking in logs, error tracking, the database and the code. The Troubleshooting Agent does that looking for you.
You describe the problem. Pluno works out which of your systems matter, gathers evidence from them, cross-checks it, and comes back with the root cause and a recommended fix, with every finding linked to its evidence.
Where you use it#
| Place | How |
|---|---|
| The Pluno chat | Ask a question in the Troubleshooting Agent web app |
| Your help desk | Select Investigate ticket in the Pluno sidebar on a ticket |
| Slack | Mention Pluno in a channel or thread, or message it directly |
| Your own tools | Start an investigation through the API |
See Run an investigation.
What it can look at#
Pluno uses only what you connect. Nothing is indexed in advance. It reads from your systems live, during each investigation.
- Your code, from the GitHub or GitLab repositories you select.
- Errors, logs and metrics, from tools such as Sentry or Datadog.
- Your data, from databases such as PostgreSQL, or analytics tools such as PostHog.
- Your infrastructure and incidents, from tools such as AWS or PagerDuty.
- Almost any other tool you use. The catalog holds more than 3,000 apps, including session recording tools such as FullStory or LogRocket.
- Your own internal APIs, added as custom APIs.
- Your tickets and knowledge, if your workspace has a help desk connected.
See Connect your stack.
What it does and does not do#
The Troubleshooting Agent works in an isolated, temporary workspace. Beyond finding the cause, it helps you act on it:
- Fixes the customer can apply are posted on the ticket as an internal note, ready for your agent to pass on.
- Confirmed bugs are escalated to engineering with the recommended fix attached.
- Pull requests are opened with the proposed code change, but only after you approve.
- Incident alerts are raised when a spike in tickets lines up with a spike in errors.
It never replies to your customers itself.
Always check a finding before you act on it. Pluno shows the evidence it used so that you can.
Get started#
- Open the Troubleshooting Agent. If you already use Pluno, open Troubleshooting Agent in the Pluno web app. It uses your existing workspace and members. If you are new, sign up and create a workspace.
- Add integrations. Connect at least one system. See Connect your stack.
- Connect Slack (optional). See Use the Troubleshooting Agent in Slack.
- Open the chat and ask your first question.
Some workspaces are first asked to accept the Data Processing Agreement and the list of subprocessors. An Admin does this once.
Who can do what#
| Role | Can |
|---|---|
| Admin | Everything: connect integrations and Slack, change settings, manage the API key, the plan and usage |
| Support Agent | Run investigations in the chat and in the help desk sidebar, and open shared chats |
People who use Pluno in Slack do not need a Pluno account.
Credits#
The Troubleshooting Agent runs on credits. Every workspace starts with 4,080 free credits, and an investigation uses about 80. See Settings, plans and credits.
Last updated