3 ================================
4 PSI - Pressure Stall Information
5 ================================
8 :Author: Johannes Weiner <hannes@cmpxchg.org>
10 When CPU, memory or IO devices are contended, workloads experience
11 latency spikes, throughput losses, and run the risk of OOM kills.
13 Without an accurate measure of such contention, users are forced to
14 either play it safe and under-utilize their hardware resources, or
15 roll the dice and frequently suffer the disruptions resulting from
18 The psi feature identifies and quantifies the disruptions caused by
19 such resource crunches and the time impact it has on complex workloads
20 or even entire systems.
22 Having an accurate measure of productivity losses caused by resource
23 scarcity aids users in sizing workloads to hardware--or provisioning
24 hardware according to workload demand.
26 As psi aggregates this information in realtime, systems can be managed
27 dynamically using techniques such as load shedding, migrating jobs to
28 other systems or data centers, or strategically pausing or killing low
29 priority or restartable batch jobs.
31 This allows maximizing hardware utilization without sacrificing
32 workload health or risking major disruptions such as OOM kills.
37 Pressure information for each resource is exported through the
38 respective file in /proc/pressure/ -- cpu, memory, and io.
40 The format is as such::
42 some avg10=0.00 avg60=0.00 avg300=0.00 total=0
43 full avg10=0.00 avg60=0.00 avg300=0.00 total=0
45 The "some" line indicates the share of time in which at least some
46 tasks are stalled on a given resource.
48 The "full" line indicates the share of time in which all non-idle
49 tasks are stalled on a given resource simultaneously. In this state
50 actual CPU cycles are going to waste, and a workload that spends
51 extended time in this state is considered to be thrashing. This has
52 severe impact on performance, and it's useful to distinguish this
53 situation from a state where some tasks are stalled but the CPU is
54 still doing productive work. As such, time spent in this subset of the
55 stall state is tracked separately and exported in the "full" averages.
57 CPU full is undefined at the system level, but has been reported
58 since 5.13, so it is set to zero for backward compatibility.
60 The ratios (in %) are tracked as recent trends over ten, sixty, and
61 three hundred second windows, which gives insight into short term events
62 as well as medium and long term trends. The total absolute stall time
63 (in us) is tracked and exported as well, to allow detection of latency
64 spikes which wouldn't necessarily make a dent in the time averages,
65 or to average trends over custom time frames.
67 Monitoring for pressure thresholds
68 ==================================
70 Users can register triggers and use poll() to be woken up when resource
71 pressure exceeds certain thresholds.
73 A trigger describes the maximum cumulative stall time over a specific
74 time window, e.g. 100ms of total stall time within any 500ms window to
75 generate a wakeup event.
77 To register a trigger user has to open psi interface file under
78 /proc/pressure/ representing the resource to be monitored and write the
79 desired threshold and time window. The open file descriptor should be
80 used to wait for trigger events using select(), poll() or epoll().
81 The following format is used::
83 <some|full> <stall amount in us> <time window in us>
85 For example writing "some 150000 1000000" into /proc/pressure/memory
86 would add 150ms threshold for partial memory stall measured within
87 1sec time window. Writing "full 50000 1000000" into /proc/pressure/io
88 would add 50ms threshold for full io stall measured within 1sec time window.
90 Triggers can be set on more than one psi metric and more than one trigger
91 for the same psi metric can be specified. However for each trigger a separate
92 file descriptor is required to be able to poll it separately from others,
93 therefore for each trigger a separate open() syscall should be made even
94 when opening the same psi interface file. Write operations to a file descriptor
95 with an already existing psi trigger will fail with EBUSY.
97 Monitors activate only when system enters stall state for the monitored
98 psi metric and deactivates upon exit from the stall state. While system is
99 in the stall state psi signal growth is monitored at a rate of 10 times per
102 The kernel accepts window sizes ranging from 500ms to 10s, therefore min
103 monitoring update interval is 50ms and max is 1s. Min limit is set to
104 prevent overly frequent polling. Max limit is chosen as a high enough number
105 after which monitors are most likely not needed and psi averages can be used
108 When activated, psi monitor stays active for at least the duration of one
109 tracking window to avoid repeated activations/deactivations when system is
110 bouncing in and out of the stall state.
112 Notifications to the userspace are rate-limited to one per tracking window.
114 The trigger will de-register when the file descriptor used to define the
117 Userspace monitor usage example
118 ===============================
130 * Monitor memory partial stall with 1s tracking window size
131 * and 150ms threshold.
134 const char trig[] = "some 150000 1000000";
138 fds.fd = open("/proc/pressure/memory", O_RDWR | O_NONBLOCK);
140 printf("/proc/pressure/memory open error: %s\n",
144 fds.events = POLLPRI;
146 if (write(fds.fd, trig, strlen(trig) + 1) < 0) {
147 printf("/proc/pressure/memory write error: %s\n",
152 printf("waiting for events...\n");
154 n = poll(&fds, 1, -1);
156 printf("poll error: %s\n", strerror(errno));
159 if (fds.revents & POLLERR) {
160 printf("got POLLERR, event source is gone\n");
163 if (fds.revents & POLLPRI) {
164 printf("event triggered!\n");
166 printf("unknown event received: 0x%x\n", fds.revents);
177 In a system with a CONFIG_CGROUP=y kernel and the cgroup2 filesystem
178 mounted, pressure stall information is also tracked for tasks grouped
179 into cgroups. Each subdirectory in the cgroupfs mountpoint contains
180 cpu.pressure, memory.pressure, and io.pressure files; the format is
181 the same as the /proc/pressure/ files.
183 Per-cgroup psi monitors can be specified and used the same way as