Shulou(Shulou.com)06/02 Report--

Memory:

Container_memory_rss

RSS memory, or resident memory set (Resident Set Size), is allocated to processes to use actual physical memory rather than virtual memory cached on disk. RSS memory includes all allocated stack and heap memory, as well as the memory space occupied by shared libraries loaded into physical memory, but does not include memory that enters the swap partition.

Container_memory_usage_bytes

The amount of memory currently used, including all used memory, whether accessed or not.

Container_memory_max_usage_bytes

Record of maximum memory usage.

Container_memory_cache

Usage of cache (cache). Cache is a small but high-speed memory located between CPU and main memory, which is designed to improve the speed of data exchange between cpu and memory.

Container_memory_swap

Virtual memory usage. Virtual memory (swap) refers to the use of disks to simulate memory usage. When the physical memory is running out or reaches a certain proportion, some of the unused memory data can be exchanged to the hard disk for storage, and then transferred to the physical memory when needed.

Container_memory_working_set_bytes

Current memory working set (working set) usage.

Container_memory_max_usage_bytes (maximum available memory) > container_memory_usage_bytes (memory requested + memory used by working sets) > container_memory_working_set_bytes (working set memory) > container_memory_rss (resident memory set)

PromQL Foundation

Value

explain

Http_request_total {} instantaneous vector expression, select the latest data http_request_total {} [5m] interval vector expression, select the data within 5 minutes http_request_total {} offset 5m before the instantaneous sample data http_request_total {} [1D] offset 1d the sample data PromQL aggregation operation within the interval of the day yesterday.

Value

explain

Sum () by (cluster_name) the summation of the value values in each result in unlimited units, for example, to request a transient alarm for the first n time series of the total time topk, not to observe the n time series quantile (0.5, http_requests_total) after the graph bottomk (), when φ is 0.5. It means to find the median in the current sample data, min max avg, min, max, average, count (), count each result, and count value per unit number of times count_values (). PromQL built-in function

Value

explain

Value

explain

Rate (5m) takes the average number of increments per second over a period of time, the total increment within 5m / 5m is suitable for the slowly changing counter (counter) irate (5m) to calculate the rate, the counter (counter) increase (5m) is suitable for the total amount of time increments over a period of time, and the total increment predict_linear (node_filesystem_free {job= "node"} [1h], 4 * 3600) within 5m is suitable for gauges. Predicting the changing trend of data by and without (signature) without are used to remove enumerated tags from the calculation results while retaining other tags. By, on the other hand, retains only the tags listed in the result vector and removes the rest. Through without and by, the data can be aggregated according to the problems of the sample. Avg without (cpu) (rate (node_cpu_seconds_total {mode= "idle"} [5m])) without is not grouped by cpu tag, and then the average is calculated. Ceil (node_load5 {instance= "192.168.1.75 instance= 9100"}) rounding day_of_month () returns the day of the month in which the given UTC time is located (clamp_max (node_load5 {instance= "192.168.1.75 instance= 9100"}, 2) enter an instantaneous vector and the maximum value, and if the sample data value is greater than max, change it to max. Otherwise, the invariant changes (node_load5 {instance= "192.168.1.75 instance= 9100"} [1m]) inputs an interval vector and returns the number of changes (instantaneous vectors) of each sample data value within this interval vector, if the sample data value does not change. The result is 1delta (cpu_temp_celsius {host= "zeus"} [2h]), which calculates the difference between the first element and the last element of an interval vector v and is used in time series of type Gauge.

Welcome to subscribe "Shulou Technology Information " to get latest news, interesting things and hot topics in the IT industry, and controls the hottest and latest Internet news, technology news and IT industry trends.