Monitor a database

Two tabs on a database’s details page show how it is doing. Nodes & Monitoring shows the database’s node with its latest CPU, memory, and storage readings, and opens charts over time. Activity Timeline records everything done to the database, by people and by TIR.

Before you begin

• A database that has finished its first launch. While a new database is Creating, and after a launch that failed, its page shows the launch progress instead of the tabs. See Create a database.
• The DBaaS read permission in the project. See TIR Access Control.

Check the node’s readings

1. In the side menu, under Databases, select DBaaS, then select the database’s name.
2. Select the Nodes & Monitoring tab.

While the database is Running, the tab shows its node, the primary, with the latest CPU, memory, and disk use as percentages, and the node’s VPC address when a VPC is attached. While the database is Scaling for a storage increase, the node drops off the tab and comes back when scaling finishes. The readings, the charts, and the activity timeline below all refresh on their own while the browser tab is in view.

The Overview tab also shows how full the disk is, under Storage in Plan & Usage. Once 80% of the disk is in use, Storage offers Increase storage while the database is Running. The disk holds the operating system as well as your data. To add storage, see Edit a database.

Open the charts

To open a node’s charts, select it, or choose View Monitoring in its Actions menu. The charts show that node’s own figures, while the database is Running.

Chart	What it shows	Lines
CPU Utilization	vCPU usage as a share of the plan’s cores	One
Memory Utilization	RAM usage as a share of the plan’s memory	One
Storage Utilization	Disk space used as a share of the disk’s size	One
Disk Latency	The average time per read and per write, in ms	Read, Write
Disk IOPS	Read and write operations per second	Read, Write
Disk Throughput	Bytes read and written per second	Read, Write
Network	Bytes received and sent per second	In, Out

Choose a time range of 1h, the default, 6h, 1d, 7d, or 30d. Longer ranges have points further apart: 1 minute apart on 1h, 5 minutes on 6h, 15 minutes on 1d, 1 hour on 7d, and 4 hours on 30d.

Under each chart, Avg, Peak, and Min give the average, highest, and lowest value of each line over the range. Percentage charts run from 0 to 100%. Throughput and network read in B/s, KB/s, MB/s, or GB/s, whichever suits the largest value, where 1 KB is 1,024 bytes.

If a chart has only one point for the range, choose a shorter range.

The charts cover only the node’s resources; connection counts, slow queries, and the database server logs fall outside them.

Read the activity timeline

Select the Activity Timeline tab. It lists everything done to the database since it was created, newest first and grouped by day. Once the database has finished its first launch, the tab is available in every status, including Stopped. It holds the newest 5,000 entries at most.

Each entry gives the time, on a 24-hour clock in your local time, what happened, any details, and who acted: a person’s email address, TIR for what TIR does on its own, or Snapshot schedule for scheduled snapshots. A person whose account has since been deleted is shown as a dash.

To narrow the list, choose Manual for actions people took, or Automatic for actions TIR took, or All. Search activity finds entries by their title, detail, size change, or who acted.

What people do is recorded under these titles:

• The database: Database created, Database created (restored from snapshot), Database started, Database stopped, Database restarted, Database scaled, and Database deleted.
• Snapshots: Snapshot created, Snapshot deleted, Snapshot restored, Snapshot schedule configured, and Snapshot schedule disabled.
• Settings: Database user password reset, Parameter group attached, Parameter group detached, Extensions added, and IP allowlist updated.

What TIR does on its own is recorded as Scheduled snapshot created, Snapshot completed, Snapshot failed, and Scaling did not complete.

The detail line fills in the specifics. A password reset names the user, a parameter group entry names the group, and “Extensions added” names the extensions. An allowlist entry says how many address ranges are allowed, such as “2 address ranges allowed”. A schedule entry gives the schedule and how many snapshots it keeps. A restore is recorded on both databases. The source shows “Snapshot restored” with the snapshot and the new database’s name. The new database’s entry names the source and the snapshot.

A deleted database’s “Database deleted” entry can be read while the database is Terminating, or when it shows Failed after a failed delete.

Export the activity

To keep a copy, choose Export, then Download CSV or Download JSON. The file holds the entries shown at that moment, after any filter and search, and is named after the database, such as mydb-activity.csv.

Each exported entry has the fields time, event, detail, change, by, and outcome. Outcome is ok, attention, or failed.

For help with empty charts or tables, see Troubleshooting TIR DBaaS. To keep a copy of the data itself, see Take a snapshot of a database.