Next Compiling multiple JMeter runs into a single chart and report
JMeter is a powerful testing tool. R can be used to process the raw test logs from JMeter and generate the final product, Analysis and Reports.
The JMeter gui should only be used for design and dry run the test set. Actual test results should be taken from JMeter runs in headless mode.
jmeter -n -t "mytestset.jmx" -l "myapp_v0.1_results.csv"tr <- list() # Test Report
tr$folder <- "/folder/containing/csv/data"tr$current$contents <- read_csv(paste(tr$folder,"/myapp_v0.1_results.csv", sep=""))JMeter CSV Log Format
# pivot the table so we can see all the column headers in a column
tibble(columnHeaders=colnames(tr$current$contents),
exampleValue=unlist(tr$current$contents[1,], use.names = F)) %>%
knitr::kable()| columnHeaders | exampleValue |
|---|---|
| timeStamp | 1593897999747 |
| elapsed | 178 |
| label | homepage |
| responseCode | 200 |
| responseMessage | OK |
| threadName | Stepping Thread Group - capacity testing 1-1 |
| dataType | text |
| success | TRUE |
| failureMessage | NA |
| bytes | 1133 |
| sentBytes | 296 |
| grpThreads | 1 |
| allThreads | 1 |
| URL | http://localhost:5000/ |
| Latency | 173 |
| IdleTime | 0 |
| Connect | 106 |
This table shows a attribute called “elapsed”, with an example value of 178 which is in milliseconds (1000 ms = 1 sec).
Aggregate and Sort
tr$current$summary <-
tr$current$contents %>%
select(Endpoint = label, elapsed = elapsed) %>%
group_by(Endpoint) %>%
summarise(Average=mean(elapsed)) %>%
arrange(desc(Average))Plot
positions = rev(tr$current$summary$Endpoint)
ggplot(data = tr$current$summary, aes(x=Endpoint, y=Average)) +
geom_bar(stat="identity") +
scale_x_discrete(limits = positions) +
coord_flip() 