---
title: Service operations
description: "Read the history of a database service and use Summary, Output, Errors and Inventory to inspect an operation run."
search:
  keywords: ["operation history", "Ansible", "failed steps", "inventory", "deploy logs"]
---

The **Operations** tab of a database service shows its operation history,
including deploys, configuration changes, backups and upgrades. Each row is
one run for the selected service.

The console also has a separate **Operations** page in its main navigation.
That page lists runs for all database services and DCS clusters. The
**Operations** tab described here lists only runs for the service you have
open.

A viewer can read operations and their logs. For the logs of PostgreSQL,
Patroni, PgBouncer or the operating system, use
[Logs & Diagnostics](/docs/v2026.11.12-beta/logs-and-support) instead.

## Open a run

The history table is at the top of the tab. Find the run by its operation,
status and start time. Each row also shows its duration and finish time.
The search box above the table, on the right, filters only the current
page. Use the page controls below the table to browse older runs.

Select **View logs** near the right end of a row. A panel opens below the
table with **Summary** selected. The row's button changes to **Close**.

The panel header shows the operation name and run ID on the left, with the
status and download button on the right. Below the header are four tabs:
**Summary**, **Output**, **Errors** and **Inventory**. They all belong to
the selected run.

Some operations run no Ansible playbook. For example, dropping an extension
runs directly against PostgreSQL. The panel then says "This operation runs
no playbook, so it has no log."

## Summary

Start here to follow a run. Summary shows the Ansible automation steps in
order, grouped under play headings. Each step has a result on the left and
host counts and timing on the right.

Select the chevron at the left of a step to expand its per-host results.
Error messages appear beneath the affected host. A skipped step can show
why it was skipped. The **Search steps** box above the list searches step
names, hosts and error messages.

When Ansible reports its final recap, counts such as **ok**, **changed**,
**skipped**, **failed** and **unreachable** appear beside the run ID in the
panel header. The recap also appears on Output and Errors.

## Output

Output shows the detailed Ansible event log used to build Summary. Events
use JSON Lines format: one JSON object per line. With verbose logging
enabled, plain-text diagnostic lines can appear between events. Open this
tab when you need more detail than the step list provides.

Use **Search output** above the log to filter its lines. The controls at
the top right of the panel let you pretty-print the JSON and wrap long
lines. They change only the display; downloads keep the original log.

Summary, Output and Errors update as output arrives during a run. You can
scroll up within the panel to read earlier output. Select **Jump to latest**
to follow new output again.

## Errors

Errors shows the steps with failures, including hosts that could not be
reached. It uses the same expandable rows as Summary. An automation runtime
error also appears here when the run records one.

To investigate a failure:

1. Select **Errors** in the tab row below the panel header.
2. Expand a failed step using the chevron on its left. Read the error
   message beneath the affected host.
3. Switch to **Output** for the detailed events, or **Inventory** to check
   the target hosts and input settings.

If Errors is empty, clear **Search steps** and check the run's status in
the panel header. A run can stop before it records a failed step, so an
empty list alone does not show that it succeeded.

A **Warning** status means the operation completed but recorded a problem
that needs attention. For example, a backup can finish while the later
removal of expired backups fails.

## Inventory

Inventory shows the inputs supplied to Ansible for this run: the hosts,
host groups and configuration variables. Credential values are hidden.
Use it to check which servers the run targeted and which settings it used.
You can copy the inventory from this tab.

These inputs are a snapshot from the run. They do not update as servers or
settings change, and they do not show which resources the operation created
or the resulting server configuration. The recap counts from the other
tabs do not appear here.

## Download a log or inventory

Select the down-arrow button at the top right of the run panel, beside the
status. The selected tab determines which file you get:

| Selected tab | Downloaded file |
| --- | --- |
| Summary or Output | The detailed event log, with JSON Lines events and any plain-text diagnostics. |
| Errors | The raw Ansible error and warning log (stderr). |
| Inventory | The inventory JSON for the run, with credential values hidden. |

The Errors download contains different information from the step list you
see on screen. Ansible writes task failures to the event log. To save those
failures, download from Summary or Output.

To save the event log without opening the panel, open the three-dot menu
at the far right of the operation row and select **Download logs**.

For a report to send to support, see
[Support reports](/docs/v2026.11.12-beta/logs-and-support#support-reports).
