I was taking a quick look to K6 from loadimpact.
The graphs that I got so far show TPS, Response Time, Error rates at the global level and that is not too useful.
When I load test, I rather have those stats at the global level, but also at the flow level or at the APi level. This way if for example if I see some high latency I can tell right away if is caused by a single API or if all APIs are slow.
Or I can tell of a given API is giving say HTTP/500 or several different APIs are.
Can K6 show stats like TPS, Response Time, HTTP status at the API level, the flow level and global level?
Thanks
Yes, it can, and you have 3 options here in terms of result presentation (all involve using custom metrics to some degree):
End of test summary printed to stdout.
You output the result data to InfluxDB+Grafana.
You output the result data to Load Impact Insights.
Global stats you get with all three above, and per API endpoint stats you get out of the box with 2) and 3), but to get stats at the flow level you'd need to create custom metrics which works with all three options above. So something like this:
import http from "k6/http";
import { Trend, Rate } from "k6/metrics";
import { group, sleep } from "k6";
export let options = {
stages: [
{ target: 10, duration: "30s" }
]
};
var flow1RespTime = new Trend("flow_1_resp_time");
var flow1TPS = new Rate("flow_1_tps");
var flow1FailureRate = new Rate("flow_1_failure_rate");
export default function() {
group("Flow 1", function() {
let res = http.get("https://test.loadimpact.com/");
flow1RespTime.add(res.timings.duration);
flow1TPS.add(1);
flow1FailureRate.add(res.status == 0 || res.status > 399);
});
sleep(3);
};
This would expand the end of test summary stats printed to stdout to include the custom metrics:
Related
I use ClearML to track my tensorboard logs (from PyTorch Lightning) during training.
At a point later I start another script which connects to existing task and do some testing.
But unfortenautly I do not have all information in the second script, so I want to query them from the logged values from ClearML server.
How would I do this?
I thought about something like this, but havn't found anything in documentation:
task = Task.init(project_name="Project", task_name="name", reuse_last_task_id="Task_id, continue_last_task=True)
x_value, y_value = task.get_value(key="val/acc", mode="max")
x_value2, y_value2 = task.get_value(key="epoch", mode="x", x=x_value)
x_value would be my epoch or global step
y_value the maximum value of plot "val/acc"
x_value2 would be my epoch or global step
y_value2 the value of plot "epoch" at x_value
Disclaimer I'm part of the ClearML (formerly Trains) team.
To get an existing Task object for a running (or completed/failed) experiment, assuming we know Task ID:
another_task = Task.get_task(task_id='aabbcc')
If we only know the Task project/name
another_task = Task.get_task(project_name='the project', task_name='the name')
Notice that if you have multiple task under the same name it will return the most updated one.
Once we have the Task object, we can do:
latest_scalar_values_dict = another_task.get_last_scalar_metrics()
Which would return all the scalars min/maxm/last values, for example:
latest_scalar_values_dict = {
'title': {
'series': {
'last': 0.5,
'min': 0.1,
'max': 0.9
}
}
}
documentation here
If you need to get the entire graphs you can use task.get_reported_scalars() see docs
We are using Prometheus and Grafana for our monitoring and we have a panel for response time however I noticed after while the metrics are missing and there are a lots of gap in the panel (only for response time panel) and they comeback as soon as I restart the app (redeploying it in openshift). the service has been written in Go and the logic for the gathering response time is quite simple.
we declared the metric
var (
responseTime = promauto.NewSummaryVec(prometheus.SummaryOpts{
Namespace: "app",
Subsystem: "rest",
Name: "response_time",
}, []string{
"path",
"code",
"method",
})
)
and fill it in our handler
func handler(.......) {
start := time.Now()
// do stuff
....
code := "200"
path := r.URL.Path
method := r.Method
elapsed := float64(time.Since(start)) / float64(time.Second)
responseTime.WithLabelValues(path, code, method).Observe(elapsed)
}
and query in the Grafana panel is like:
sum(rate(app_rest_response_time_sum{path='/v4/content'}[5m]) /
rate(app_rest_response_time_count{path='/v4/content'}[5m])) by (path)
but the result is like this!!
can anyone explain what do we do wrong or how to fix this issue? is it possible that we facing some kind of overflow issue (the average RPS is about 250)? I'm suspecting this because this happen more often to the routes with higher RPS and response time!
Prometheus records the metrics continuously normally and if you query it, it returns all the metrics it collected for the time you queried.
If there is no metric when you query, that has typically three reasons:
the metric was not there (it happens when the instance restarts and you have a dynamic set of labels and there was no request yet for the label value you queried (in your case there was no query for path='/v4/content'). In such case you should see other metrics of the same job (at least up).
Prometheus had problems storing the metrics. (see the log files of prometheus for that timeframe).
Prometheus was down for that timeframe and therefore did not collect any metrics. (In that case you should have no metrics at all for that timeframe.
We retrieve information from Elasticsearch 2.7.0 and we allow the user to go through the results. When the user requests a high page number we get the following error message:
Result window is too large, from + size must be less than or equal to:
[10000] but was [10020]. See the scroll api for a more efficient way
to request large data sets. This limit can be set by changing the
[index.max_result_window] index level parameter
The thing is we use pagination in our requests so I don't see why we get this error:
#Autowired
private ElasticsearchOperations elasticsearchTemplate;
...
elasticsearchTemplate.queryForPage(buildQuery(query, pageable), Document.class);
...
private NativeSearchQuery buildQuery() {
BoolQueryBuilder boolQueryBuilder = QueryBuilders.boolQuery();
boolQueryBuilder.should(QueryBuilders.boolQuery().must(QueryBuilders.termQuery(term, query.toUpperCase())));
NativeSearchQueryBuilder nativeSearchQueryBuilder = new NativeSearchQueryBuilder().withIndices(DOC_INDICE_NAME)
.withTypes(indexType)
.withQuery(boolQueryBuilder)
.withPageable(pageable);
return nativeSearchQueryBuilder.build();
}
I don't understand the error because we retreive pageable.size (20 elements) everytime... Do you have any idea why we get this?
Unfortunately, Spring data elasticsearch even when paging results searchs for a much larger result window in the elasticsearch. So you have two options, the first is to change the value of this parameter.
The second is to use the scan / scroll API, however, as far as I understand, in this case the pagination is done manually, as it is used for infinite sequential reading (like scrolling your mouse).
A sample:
List<Pessoa> allItens = new ArrayList<>();
String scrollId = elasticsearchTemplate.scan(build, 1000, false, Pessoa.class);
Page<Pessoa> page = elasticsearchTemplate.scroll(scrollId, 5000L, Pessoa.class);
while (true) {
if (!page.hasContent()) {
break;
}
allItens.addAll(page.getContent());
page = elasticsearchTemplate.scroll(scrollId, 5000L, Pessoa.class);
}
This code, shows you how to read ALL the data from your index, you have to get the requested page inside scrolling.
Consider an example of testing API's with Gatling. For some weird requirement i had to get a scenario for each user
var scenarioList // This is of type mutable list
I have plenty of scenarios added to this list as my request body should differ for each user or the request won't be processed.This individual scenarios have following gatling simulation configured at this moment
Ex: scenarioList += scenario1. inject(rampUsers(1) over (1 minutes)
scenarioList += scenario2. inject(rampUsers(1) over (1 minutes)
scenarioList += scenario3. inject(rampUsers(1) over (1 minutes)
.
.
.
so on
Now in the global setup as below while calling all these scenarios
setUp(scenarioList: _*).assertions(
forAll.successfulRequests.percent.gte(90)
)
Suppose i have 1000 users (scenarioList size is 1000), The problem here would be all of the 1000 users would start at the same time but i want to ramp up these users. So the question comes of ramping up the scenarios instead of running them parallely.
Is this possible ? If not is there any other approach to follow ?
I can't have the luxury of running the same scenario with multiple users as the body of the requests change. Please let me know.
I was able to solve this problem by using feeders within the scenario so i don't need to create multiple scenarios.
With feeders Gatling provides option to parameterize your request body of any http request.
Code Example:
var randomSession = Iterator.continually(Map("randsession" -> ( req.replace("0000000000", randomStringGenerator.randomString(10)))))
val httpConf = http
.baseURL("http://localhost:5000")
.acceptHeader("text/html,application/xhtml+xml,application/xml;q=0.9,*/*;q=0.8")
.userAgentHeader("Mozilla/4.0(compatible;IE;GACv10. 0. 0. 1)")
val scn = scenario("Activate")
.feed(randomSession)
.exec(http("activate request")
.post("/login/activate")
.body(StringBody("""${randsession}"""))
.check(status.is(200)))
.pause(5)
setUp(
scn.inject(atOnceUsers(5))
).protocols(httpConf)
}
I want to use amazon Dynamo DB with rails.But I have not found a way to implement pagination.
I will use AWS::Record::HashModel as ORM.
This ORM supports limits like this:
People.limit(10).each {|person| ... }
But I could not figured out how to implement following MySql query in Dynamo DB.
SELECT *
FROM `People`
LIMIT 1 , 30
You issue queries using LIMIT. If the subset returned does not contain the full table, a LastEvaluatedKey value is returned. You use this value as the ExclusiveStartKey in the next query. And so on...
From the DynamoDB Developer Guide.
You can provide 'page-size' in you query to set the result set size.
The response of DynamoDB contains 'LastEvaluatedKey' which will indicate the last key as per the page size. If response does't contain 'LastEvaluatedKey' it means there are no results left to fetch.
Use the 'LastEvaluatedKey' as 'ExclusiveStartKey' while fetching next time.
I hope this helps.
DynamoDB Pagination
Here's a simple copy-paste-run proof of concept (Node.js) for stateless forward/reverse navigation with dynamodb. In summary; each response includes the navigation history, allowing user to explicitly and consistently request either the next or previous page (while next/prev params exist):
GET /accounts -> first page
GET /accounts?next=A3r0ijKJ8 -> next page
GET /accounts?prev=R4tY69kUI -> previous page
Considerations:
If your ids are large and/or users might do a lot of navigation, then the potential size of the next/prev params might become too large.
Yes you do have to store the entire reverse path - if you only store the previous page marker (per some other answers) you will only be able to go back one page.
It won't handle changing pageSize midway, consider baking pageSize into the next/prev value.
base64 encode the next/prev values, and you could also encrypt.
Scans are inefficient, while this suited my current requirement it won't suit all!
// demo.js
const mockTable = [1,2,3,4,5,6,7,8,9,10,11,12,13,14,15,16,17,18,19,20]
const getPagedItems = (pageSize = 5, cursor = {}) => {
// Parse cursor
const keys = cursor.next || cursor.prev || [] // fwd first
let key = keys[keys.length-1] || null // eg ddb's PK
// Mock query (mimic dynamodb response)
const Items = mockTable.slice(parseInt(key) || 0, pageSize+key)
const LastEvaluatedKey = Items[Items.length-1] < mockTable.length
? Items[Items.length-1] : null
// Build response
const res = {items:Items}
if (keys.length > 0) // add reverse nav keys (if any)
res.prev = keys.slice(0, keys.length-1)
if (LastEvaluatedKey) // add forward nav keys (if any)
res.next = [...keys, LastEvaluatedKey]
return res
}
// Run test ------------------------------------
const runTest = () => {
const PAGE_SIZE = 6
let x = {}, i = 0
// Page to end
while (i == 0 || x.next) {
x = getPagedItems(PAGE_SIZE, {next:x.next})
console.log(`Page ${++i}: `, x.items)
}
// Page back to start
while (x.prev) {
x = getPagedItems(PAGE_SIZE, {prev:x.prev})
console.log(`Page ${--i}: `, x.items)
}
}
runTest()
I faced a similar problem.
The generic pagination approach is, use "start index" or "start page" and the "page length".
The "ExclusiveStartKey" and "LastEvaluatedKey" based approach is very DynamoDB specific.
I feel this DynamoDB specific implementation of pagination should be hidden from the API client/UI.
Also in case, the application is serverless, using service like Lambda, it will be not be possible to maintain the state on the server. The other side is the client implementation will become very complex.
I came with a different approach, which I think is generic ( and not specific to DynamoDB)
When the API client specifies the start index, fetch all the keys from
the table and store it into an array.
Find out the key for the start index from the array, which is
specified by the client.
Make use of the ExclusiveStartKey and fetch the number of records, as
specified in the page length.
If the start index parameter is not present, the above steps are not
needed, we don't need to specify the ExclusiveStartKey in the scan
operation.
This solution has some drawbacks -
We will need to fetch all the keys when the user needs pagination with
start index.
We will need additional memory to store the Ids and the indexes.
Additional database scan operations ( one or multiple to fetch the
keys )
But I feel this will be very easy approach for the clients, which are using our APIs. The backward scan will work seamlessly. If the user wants to see "nth" page, this will be possible.
In fact I faced the same problem and I noticed that LastEvaluatedKey and ExclusiveStartKey are not working well especially when using Scan So I solved Like this.
GET/?page_no=1&page_size=10 =====> first page
response will contain count of records and first 10 records
retry and increase number of page until all record come.
Code is below
PS: I am using python
first_index = ((page_no-1)*page_size)
second_index = (page_no*page_size)
if (second_index > len(response['Items'])):
second_index = len(response['Items'])
return {
'statusCode': 200,
'count': response['Count'],
'response': response['Items'][first_index:second_index]
}