Tiering monitoring

Volumes with an upper (autotiering) tier evacuate the coldest data down to the lower tier once the upper tier's pinned size is exceeded — see the Tiering chapter for how tiers are configured. These three views let you watch that evacuation happen live, review how well each tier has been performing, and see how full each tier's upper tier has been running.

If no volumes in the system have an upper tier configured, these commands report that tiered volumes aren't configured rather than showing empty data.

Live evacuation

monitoring tiers live streams evacuation activity — how much data is being moved down to the lower tier, and read IOPS — for every tiered volume, once a second.

Using the web interface

Open Monitoring > Tiering and select the live view.

Using the CLI

monitoring tiers live [--interval=<interval>] [--interactiveChart] [--iterations=<numbers>] [--singleSample]

Option	Description
--interval=<interval>	The sampling interval in seconds.
--interactiveChart	Show a chart instead of textual information. Windows CLI only.
--iterations=<numbers>	Stop automatically after this many samples. Once selected, you can't exit before reaching that number.
--singleSample	Get a single sample instead of monitoring continuously.

Examples

Continuously monitor evacuation activity:

monitoring tiers live

Collect exactly 20 samples and stop automatically:

monitoring tiers live --iterations 20

Tier statistics

monitoring tiers stats is a per-tier report card: for each tiered volume, it shows how long recent evacuations took, the average latency during evacuation versus outside of it, how much of the tier's reads were served from the upper tier versus the lower tier (and the average I/O size of those lower-tier reads), whether the tier is currently configured to be used for writes only, and the timestamps of the last evacuation cycle (when evacuation became a candidate, when it started, and when it last completed). It also includes a set of normalized scores (usage, read distribution, evacuation duration, and latency impact) that summarize how well the tier is performing overall. It takes no parameters — it always reports every tiered volume.

Using the web interface

Open Monitoring > Tiering and select the statistics view.

Using the CLI

monitoring tiers stats

Example

monitoring tiers stats

Tier usage distribution

monitoring tiers usage_distribution shows, for each tiered volume, the percentage of time its upper tier has spent at different fill levels relative to its pinned size: comfortably below it, above it, or running hot enough that evacuation is working hard to keep up — plus how much time, if any, was spent bypassing the upper tier entirely (writes going straight to the lower tier because the upper tier was full). This is the view to check if you're trying to decide whether a volume's upper tier is sized correctly — a volume that spends a lot of time in the high-consumption or bypass state would benefit from a larger pinned size. It takes no parameters.

Using the web interface

Open Monitoring > Tiering and select the usage distribution view.

Using the CLI

monitoring tiers usage_distribution

Example

monitoring tiers usage_distribution

---

See also: Health and capacity · Live monitoring · Historical monitoring