We have multiple instance of a micro service behind the Kong API gateway where we want to balance the load for the user requests.
Say Micro service 1 is multiplied in multiple instances which are kept behind the KONG API gateway; in such case the request from user 1 should hit the first instance and the request from user 2 should hit some other instance of same service based on their availability (load balancing). (ie) Whether can i have multiple upstream URL for a single API in kong. we dont want to use nginx for load balancing. Please advice how can we solve it.
Ring-balancer Strategy can be used in Kong if you don't want DNS-based loadbalancing. For details please refer to Kong Load Balancing Reference!
# create an upstream
$ curl -X POST http://kong:8001/upstreams \
--data "name=address.v1.service"
# add two targets to the upstream
$ curl -X POST http://kong:8001/upstreams/address.v1.service/targets \
--data "target=192.168.34.15:80"
--data "weight=100"
$ curl -X POST http://kong:8001/upstreams/address.v1.service/targets \
--data "target=192.168.34.16:80"
--data "weight=50"
# create an API targeting the Blue upstream
$ curl -X POST http://kong:8001/apis/ \
--data "name=address-service" \
--data "hosts=address.mydomain.com" \
--data "upstream_url=http://address.v1.service/address"
Requests with host header set to address.mydomain.com will now be proxied by Kong to the two defined targets; 2/3 of the requests will go to http://192.168.34.15:80/address (weight=100), and 1/3 will go to http://192.168.34.16:80/address (weight=50).
Starting from 0.10 you will be able to create a named upstream, and associate/remove targets from it.
For example if you have upstream_url=http://helloworld/ you can create a helloworld upstream and associate targets to it:
curl -d "name=helloworld" 127.0.0.1:8001/upstreams
curl -d "host=some.host.com" 127.0.0.1:8001/upstreams/helloworld/targets/
curl -d "host=2.2.2.2" 127.0.0.1:8001/upstreams/helloworld/targets/
Related
We have a few jobs in production rundeck, due to other application running and throttling CPU, rundeck sometimes goes slow and lag sometimes nearly dead but rundeckd process will be still running. I wanted to set up cron(which I know) to query rundeck healthcheck and report if its sluggish till we migrate rundeck to dedicated VM. I found rundeck3.3 has api enabled by default and I am able to query in browser like http://rundesckhost:4440/metrics/ping which will return pong. Apparently http://rundeckhost:4440/metrics/healthcheck shows json
{"dataSource.connection.time":{"healthy":true,"message":"Datasource connection healthy with timeout 5 seconds"},"quartz.scheduler.threadPool":{"healthy":true}}
I have generated API token of admin user for authentication in scripting purpose. All I wanted to know how to use API to curl with token to get the result ? I tried following but couldn't get the required result.
curl --location --request GET 'http://rundeckhost:4440/metrics/ping' \
--header 'Accept: application/json' \
--header 'X-Rundeck-Auth-Token: <generated token here>'
My ENV spec:
OS : Centos 8 |
Java : 8 |
Rundeck: 3.3 community edition |
Scripting Language : curl, bash
Your ping API call is wrong, with the following call it works:
curl --location --request GET 'http://rundeckhost:4440/api/35/metrics/ping' \
--header 'Accept: application/json' \
--header 'X-Rundeck-Auth-Token: <your-token-here>'
Just in case, also you can use RD CLI tool for that, with the command:
$ rd metrics ping
Result:
# [2020-07-31T09:09:37.200] Pinging server...
# [2020-07-31T09:09:37.503] pong
More insight answer to my own question :
if anyone is having issues in using URL http://rundeckhost:4440/api/35/metrics/ping then it would be wrong api version which could be different for different versions of rundeck. so you can find out the right version by accessing your own rundeck http://rundeckhostIP:4440/api/ which will show up xml with api version in it as follows:
Use that version number in the querying URL in which in my case its 35.
Im trying to deploy a plan that has artefacts from external service for this I want to download via curl those files that I will pass as variables... however I am not able to set programmaticly the variables with the deploymet call
curl -k -u user:passord -X POST -d "bamboo.myVariable=someurl" BASE_BAMBOO_URL/bamboo/rest/api/latest/queue/PROJECT-ID
Trying to do the same with the deployment API fails
curl BASE_BAMBOO_URL/bamboo/rest/api/latest/deploy/project/1321123123 -u user:passord-X POST -d "bamboo.myVariable=callMEwithDATA"
Trying to add that into the API fails as does trying to pass it thru JSON
curl -X POST BASE_BAMBOO/bamboo/rest/api/latest/deploy/project/1320058 -u user:passord -H "Accepts: application/json" -H "Content-Type: application/json" -d '{"name":"release-1", "myVariable":"ARTEFACT_URL"}'
To continue with a request the variables have to be passed as query params... a sad sad reality is that the Bamboo API is very messed up
bamboourl&executeAllStages=true&bamboo.variable.MYVAR=1234
Is it possible to add a remote with GitHub API same as git remote add upstream <repository URL>? I have read the docs but I can not find the way.
ref
- How to update a fork from it's original via the Github API
Thanks in advance.
# Get original hash
curl "https://api.github.com/repos/:original/:repo/git/refs/heads/master"
# Follow upstrem change
curl -X POST \
-H "Authorization: token <TOKEN>" \
--data '{"base":"<fork/BRANCH>","head":"<original/HASH>"}' \
"https://api.github.com/repos/:fork/:repo/test_project/merges"
I have a question about how can I configure -H "Host: my.host.com" from gateway to my api.
I have the following logic:
I try to cover, for example, my-api.net via apiman-gateway.
For example, I can do the following direct call to my api like 'curl -X GET --header 'Accept: */*' -H "Host: www.google.com" 'http://my-api.net/accounts?accountNumber=12314123'.
Based on this call, my api will know, that my consumer is google for example. But once I try to make a call via apiman-gateway 'curl -X GET --header 'Accept: */*' -H "Host: www.google.com" 'http://apiman-gateway/GoogleOrganization/API/V1/accounts?accountNumber=12314123', the host is not possible to pass, and implementation of my api tells me that it does not know any host name like already deployed my-api.net. Probably someone can tell me how can I set up for example via policy, probably just to make a call from apiman-gateway to my api with the host parameters like above?
I am new to openstack, trying to use Ceilometer python API to pull some data from a testbed server, I am accessing the server from a remote site
the problem is that I cannot figure out how get the an authorization token
I used the following command
curl -i 'http://HOST:8774/' -X POST -H "Content-Type: application/json" -H "Accept: application/json" -d/tokens auth": {"tenantName": "project", "passwordCredentials": {"username": "user", "password": "password"}}}'
But it does not give me anything,
curl -X GET -H "X-Auth-Token:$MY_TOKEN" http://HOST:8774/tokens
also does not give me any token
From your use of port 8774 I suspect you might be using DevStack. Try this
curl -s -X POST http://$OPENSTACK_KEYSTONE_HOST:5000/v2.0/tokens -d '{"auth": {"passwordCredentials": {"username":"my-username", "password":"my-password"}, "tenantName":"my-tenantName"}}
In DevStack Keystone (the auth service you get tokens from) is running on port 5000 by default. This may or may not be true in your case. Ask your friendly OpenStack operator what host (and port) Keystone is running on and put that in place of $OPENSTACK_KEYSTONE_HOST:5000