Using/Verifying the Autoscale service from Apps Manager UI in 5 minutes

Recently at a customer site I was asked to show how the Autoscale service shipped by default with Pivotal Cloud Foundry would work. Here is how we demoed that in less then 5 minutes.

1. Select an application to Autoscale and click on the "Autoscaling" radio option.

2. Select "Manage Autoscaling" link as shown below.

3. Set the maximum instance limit to "4" and click Save as shown below. You can also set minimum to 1 instance if you want to which will make it easier to verify the scaling of instances as one instance can easily be put under pressure.

4. Now lets set a "Scaling Rule" by clicking on the "Edit" link as shown below.

5. Now lets add a CPU rule by clicking on the "Add" link as shown below.

6. Now define a CPU rule as shown below and click on Save. Don't forget to make it active using the radio option. In this example we use very low threshold BUT it would be better to increase this to something more realistic like 30% and 60% respectively.

Now at this point we are ready to test the Autoscale service BUT to do that we are going to have to create some load. Many different ways to do that but "ab" on my Mac was the fastest way.

8. Create some load on an endpoint for your application to force CPU utilization to increase as shown below

pasapicella@pas-macbook:~$ ab -n 10000 -c 25 http://springboot-actuator-appsmanager-delightful-jaguar.cfapps.io/employees
This is ApacheBench, Version 2.3 <$Revision: 1807734 $>
Copyright 1996 Adam Twiss, Zeus Technology Ltd, http://www.zeustech.net/
Licensed to The Apache Software Foundation, http://www.apache.org/

Benchmarking springboot-actuator-appsmanager-delightful-jaguar.cfapps.io (be patient)
Completed 1000 requests
Completed 2000 requests
Completed 3000 requests
Completed 4000 requests
Completed 5000 requests
Completed 6000 requests
Completed 7000 requests
Completed 8000 requests
Completed 9000 requests
Completed 10000 requests
Finished 10000 requests

....

9. If you return to Apps Manager UI soon enough you will see that the Autoscale service has fired events to add more instances as per the screen shots below.

It's worth noting that the CF CLI Plugin for Autoscale can also show us what we have defined as as shown below. More information on this plugin is as follows

https://docs.run.pivotal.io/appsman-services/autoscaler/using-autoscaler-cli.html#install

View which applications are using the Autoscaler service:

pasapicella@pas-macbook:~$ cf autoscaling-apps
Presenting autoscaler apps in org apples-pivotal-org / space development as papicella@pivotal.io
OK
Name                              Guid                                   Enabled   Min Instances   Max Instances
springboot-actuator-appsmanager   6c137fea-6a99-4069-8031-a2aa3978804c   true      2               4

View events for an application that has Autoscaler service bound to it:

pasapicella@pas-macbook:~$ cf autoscaling-events springboot-actuator-appsmanager
Presenting autoscaler events for app springboot-actuator-appsmanager for org apples-pivotal-org / space development as papicella@pivotal.io
OK
Time                   Description
2018-04-20T09:56:30Z   Scaled down from 3 to 2 instances. All metrics are currently below minimum thresholds.
2018-04-20T09:55:56Z   Scaled down from 4 to 3 instances. All metrics are currently below minimum thresholds.
2018-04-20T09:54:46Z   Can not scale up. At max limit of 4 instances. Current CPU of 20.75% is above upper threshold of 8.00%.
2018-04-20T09:54:11Z   Can not scale up. At max limit of 4 instances. Current CPU of 30.53% is above upper threshold of 8.00%.
2018-04-20T09:53:36Z   Can not scale up. At max limit of 4 instances. Current CPU of 32.14% is above upper threshold of 8.00%.
2018-04-20T09:53:02Z   Can not scale up. At max limit of 4 instances. Current CPU of 31.51% is above upper threshold of 8.00%.
2018-04-20T09:52:27Z   Scaled up from 3 to 4 instances. Current CPU of 19.59% is above upper threshold of 8.00%.
2018-04-20T09:51:51Z   Scaled up from 2 to 3 instances. Current CPU of 8.99% is above upper threshold of 8.00%.
2018-04-20T09:13:24Z   Scaling from 1 to 2 instances: app below minimum instance limit
2018-04-20T09:13:23Z   Enabled autoscaling.

More Information

https://docs.run.pivotal.io/appsman-services/autoscaler/using-autoscaler-cli.html#install

https://docs.run.pivotal.io/appsman-services/autoscaler/using-autoscaler.html

Spring Boot Application for Pivotal Cloud Cache Service

I previously blogged about the Pivotal Cloud Cache service in Pivotal Cloud Foundry as follows

http://theblasfrompas.blogspot.com.au/2017/04/getting-started-with-pivotal-cloud.html

During that post I promised it will follow with a Spring Boot application which would use the PCC service to show what the code would look like. That demo exists at the GitHub URL below.

https://github.com/papicella/SpringBootPCCDemo

The GitHub URL above shows how you can clone , package and then push this application to PCF using your own PCC service instance using the "Spring Cloud GemFire Connector"

More Information

Pivotal Cloud Cache Docs
http://docs.pivotal.io/p-cloud-cache/index.html

Getting Started with Pivotal Cloud Cache on Pivotal Cloud Foundry

Recently we announced the new cache service Pivotal Cloud Cache (PCC) for Pivotal Cloud Foundry (PCC). In short Pivotal Cloud Cache (PCC) is a opinionated, distributed, highly available, high speed key/value caching service. PCC can be easily horizontally scaled for capacity and performance.

In this post we will show how you would provision a service, login to the Pulse UI dashboard, connect using GFSH etc. I won't create a spring boot application to use the service at this stage BUT that will follow in a post soon enough.

Steps

1. First you will need the PCC service and if it's been installed it will look like this

2. Now let's view the current plans we have in place as shown below

pasapicella@pas-macbook:~$ cf marketplace -s p-cloudcache
Getting service plan information for service p-cloudcache as papicella@pivotal.io...
OK

service plan   description          free or paid
extra-small    Plan 1 Description   free
extra-large    Plan 5 Description   free

3. Now let's create a service as shown below

pasapicella@pas-macbook:~$ cf create-service p-cloudcache extra-small pas-pcc
Creating service instance pas-pcc in org pivot-papicella / space development as papicella@pivotal.io...
OK

Create in progress. Use 'cf services' or 'cf service pas-pcc' to check operation status.

4. At this point it will asynchronously create the GemFire cluster which is essentially what PCC is. For more Information on GemFire see the docs link here.

You can check the progress one of two ways.

1. Using Pivotal Apps manager as shown below

2. Using a command as follows

pasapicella@pas-macbook:~$ cf service pas-pcc

Service instance: pas-pcc
Service: p-cloudcache
Bound apps:
Tags:
Plan: extra-small
Description: Pivotal CloudCache offers the ability to deploy a GemFire cluster as a service in Pivotal Cloud Foundry.
Documentation url: http://docs.pivotal.io/gemfire/index.html
Dashboard: http://gemfire-yyyyy.run.pez.pivotal.io/pulse

Last Operation
Status: create in progress
Message: Instance provisioning in progress
Started: 2017-04-10T01:34:58Z
Updated: 2017-04-10T01:36:59Z

5. Once complete it will look as follows

6. Now in order to log into both GFSH and Pulse we are going to need to create a service key for the service we just created, which we do as shown below.

pasapicella@pas-macbook:~/pivotal/PCF/services/PCC$ cf create-service-key pas-pcc pas-pcc-key
Creating service key pas-pcc-key for service instance pas-pcc as papicella@pivotal.io...
OK

7. Retrieve service keys as shown below

pasapicella@pas-macbook:~$ cf service-key pas-pcc pas-pcc-key
Getting key pas-pcc-key for service instance pas-pcc as papicella@pivotal.io...

{
 "locators": [
  "0.0.0.0[55221]",
  "0.0.0.0[55221]",
  "0.0.0.0[55221]"
 ],
 "urls": {
  "gfsh": "http://gemfire-yyyy.run.pez.pivotal.io/gemfire/v1",
  "pulse": "http://gemfire-yyyy.run.pez.pivotal.io/pulse"
 },
 "users": [
  {
   "password": "password",
   "username": "developer"
  },
  {
   "password": "password",
   "username": "operator"
  }
 ]
}

8. Now lets log into Pulse. The URL is available as part of the output above

Login Page

Pulse Dashboard : You can see from the dashboard page it shows how many locators and cache server members we have as part of this default cluster

9. Now lets log into GFSH. Once again the URL is as per the output above

- First we will need to download Pivotal GemFire so we have the GFSH client, download the zip at the link below and extract to your file system

  https://network.pivotal.io/products/pivotal-gemfire

- Invoke as follows using the path to the extracted ZIP file

$GEMFIRE_HOME/bin/gfsh

pasapicella@pas-macbook:~/pivotal/software/gemfire/pivotal-gemfire-9.0.3/bin$ ./gfsh
    _________________________     __
   / _____/ ______/ ______/ /____/ /
  / /  __/ /___  /_____  / _____  /
 / /__/ / ____/  _____/ / /    / /
/______/_/      /______/_/    /_/    9.0.3

Monitor and Manage Pivotal GemFire
gfsh>connect --use-http --url=http://gemfire-yyyy.run.pez.pivotal.io/gemfire/v1 --user=operator --password=password
Successfully connected to: GemFire Manager HTTP service @ http://gemfire-yyyy.run.pez.pivotal.io/gemfire/v1

gfsh>

10. Now lets create a region which will use to store some cache data

$ create region --name=demoregion --type=PARTITION_HEAP_LRU --redundant-copies=1

  
gfsh>create region --name=demoregion --type=PARTITION_HEAP_LRU --redundant-copies=1
              Member                | Status
----------------------------------- | ---------------------------------------------------------------------
cacheserver-PCF-PEZ-Heritage-RP04-1 | Region "/demoregion" created on "cacheserver-PCF-PEZ-Heritage-RP04-1"
cacheserver-PCF-PEZ-Heritage-RP04-0 | Region "/demoregion" created on "cacheserver-PCF-PEZ-Heritage-RP04-0"
cacheserver-PCF-PEZ-Heritage-RP04-2 | Region "/demoregion" created on "cacheserver-PCF-PEZ-Heritage-RP04-2"
cacheserver-PCF-PEZ-Heritage-RP04-3 | Region "/demoregion" created on "cacheserver-PCF-PEZ-Heritage-RP04-3" 

Note: Understanding the region types you can create exist at the Pivotal GemFire docs but basically in the example above we create a partitioned region where primary and backup data is stored among the cache servers. As you can see we asked for a single backup copy of each region entry to be placed on a separate cache server itself for redundancy

http://gemfire.docs.pivotal.io/geode/developing/region_options/region_types.html#region_types

11. If we return to the Pulse Dashboard UI we will see from the "Data Browser" tab we have a region

12. Now lets just add some data , few entries which are simple String key/value pairs only

  
gfsh>put --region=/demoregion --key=1 --value="value 1"
Result      : true
Key Class   : java.lang.String
Key         : 1
Value Class : java.lang.String
Old Value   : <NULL>


gfsh>put --region=/demoregion --key=2 --value="value 2"
Result      : true
Key Class   : java.lang.String
Key         : 2
Value Class : java.lang.String
Old Value   : <NULL>


gfsh>put --region=/demoregion --key=3 --value="value 3"
Result      : true
Key Class   : java.lang.String
Key         : 3
Value Class : java.lang.String
Old Value   : <NULL>

13. Finally lets query the data we have in the cache

  
gfsh>query --query="select * from /demoregion"

Result     : true
startCount : 0
endCount   : 20
Rows       : 3

Result
-------
value 3
value 1
value 2

NEXT_STEP_NAME : END

13. We can return to Pulse and invoke the same query from the "Data Browser" tab as shown below.

Of course storing data in a cache isn't useful unless we actually have an application on PCF that can use the Cache BUT that will come in a separate post. Basically we will BIND to this service, connect as a GemFire Client using the locators we are given as part of the service key and then extract the cache data we have just created above by invoking a query.

More Information

Download PCC for PCF
https://network.pivotal.io/products/cloud-cache

Data Sheet for PCC
https://content.pivotal.io/datasheets/pivotal-cloud-cache