1. Proposed title of this feature request
Expand swap utilization metrics collection for reporting.
3. What is the nature and description of the request?
Provide swap utilization (throughput/latency) statistics (minimally swap-in and swap-out, as shown by vmstat), but ideally read/write totals and time for swap device too (as shown by iostat).
4. Why does the customer need this? (List the business requirements here)
* Customer is trying to understand total impact of memory overcommit in highly-dense RHEV VDI environment. Watching output of vmstat is helpful in realtime, but goal is to report on swap usage over time, as workloads vary.
5. How would the customer like to achieve this? (List the functional requirements here)
* Data should be collected and provided in ovirt_engine_history.
6. For each functional requirement listed, specify how Red Hat and the customer can test to confirm the requirement is successfully implemented.
* Success is defined as being able to create an ad-hoc report that shows the swap-in/swap-out statistics over time.
7. Is there already an existing RFE upstream or in Red Hat Bugzilla?
8. Does the customer have any specific timeline dependencies and which release would they like to target (i.e. RHEL5, RHEL6)?
* Ideally as an update to RHEV 3.3, but will accept whatever is possible.
9. Is the sales team involved in this request and do they have any additional input?
* Yes - this RFE is created by SA.
10. List any affected packages or components.
* Unknown exactly where this needs to be implemented within RHEV
11. Would the customer be able to assist in testing this functionality if implemented?
Removing from 3.6 due to capacity
Can you provide a more precise definition of the needed information? Is there a utility that provides it, which we can expose on our API?
Request is to capture swap activity as currently reported by vmstat (swap-in (si) and swap-out (so)) and iostat (tps, r/s, w/s, r, w). These two tools are currently used to monitor swap activity in realtime, customer would like to record and track this data for reporting and analysis.
This bug did not make it in time for 3.6 release, moving out
Update from the customer with additional details for how they hope to use this reporting:
We will also need to be able to check the health of the system from these new metrics. Ideally, alerts should be sent if possible. Otherwise as a step by step process for interpreting the reports should be provided so IBM is able to determine when action is required,
Shirly - can you provide a devel-ack for this for 4.2?
The new metrics store should include these statistics.
swap - used / free / cached
swap_io - in / out
Should this be considered also for VMs or only for hosts and engine machine?
(In reply to Lukas Svaty from comment #13)
> Should this be considered also for VMs or only for hosts and engine machine?
Only hosts and engine machine.
comparable to free -m | grep Swap
IO is reported (mostly 0, swapoff / swapon is causing is to rise as expected)
available and correctly reported
verified in ovirt-engine-metrics-1.1.1-1.el7ev.noarch
Since the problem described in this bug report should be
resolved in a recent advisory, it has been closed with a
resolution of ERRATA.
For information on the advisory, and where to find the updated
files, follow the link below.
If the solution does not work for you, open a new bug report.