Admin health-check

Admin health-check

It is the responsibility of the Admin

1)In tools>operations > Dashboard > health reports > in that we have all monitoring agents.

if you go to health reports we can see many monitoring check like:

CQ Dispatcher configuration

Example content packages

CQ HTML library Manager config

Replication and Transport Users

Sling Java Script Handler

Sling JSP script handler

Sling Referrer filter

User Profile Default Access

Sling Get Servlet

Default login accounts

CRXDE support

Authorizable Node Name Generation

WCM  Filter Config

WebDev Access

Web Server Configuration

Note:Security check are explained separately.

for example a small scenario:

any page that is taking longtime to load.

a)we need to monitor which are the pages??

we need to see top 10 pages of the responsive time and mail them  to the Dev team.

for any page we  can set the threshold, for example if i set the threshold as the 10ms for the page, any page the respond time above the threshold level is considered as the critical level.

you can right click above top mentioned link, which starts with the http://xxxxxxxxxxxxxxxxxxxxxx

which provides the complete page list.

———————————————————————————————————————————————————

How to trace the top 10 pages of the request which are having slow response??

—————————————————————————————————————————————————————

Thses are basically from the Health Reports:

Systemchecks:

Maintenance Task revisionCleanupTask

Maintenance task workflowpurgetask

(all the workflows has to be cleared which is in archive state, with the help of workflowpurgetask)

Maintenance task com.day.cq.audit.implement.

(also we need to keep on clearing the audit, each time you perform the activity, the var audit is getting full, which has to be cleanup).

what are all come under audit??

whenever the activity is performed,automatically the workflow and audit log will get filled, we need to cleanup them,

Replication queues:

for example there is a connectivity problem between author and publish that is replication agent problem because of some latency issues i lost the connectivity.

Whenever i activate the pages it goes in to the replication queue, if the publish is down automatically the replication queue also become down.

The moment you see the replication queue is critical there is a worry for you.if there is no critical that means it got transported.

note:rishi is having compressive maintanence sheet with him.

—————————————————————————————————————————————————————

Whenever you have a search engine, you will have a query that is running, you need to ensure that optimized queries are running.

How to ensure that, you can go and look at  the performance of queries??

we can see in the query performance  related to the health check.

whenever the query time is too much then regular expected it is just shown in the critical.

whenever the application is running slowly just go and see all the stakeholders related to that application.

—————————————————————————————————————————————————————

The issue we need to get worried:

For example:my server has got down

1)Disk space issue

2)Out of memory issue.

whenever anyone is analyzing we need to provide them with thread dump& Heap dump which is adobe related.

How to get Thread dump & Heap dump??

there is an option diagnosis > there we have two options > Download Thread Dump & Heap dump.

or 

we can get this with the help of OSGI system console also.

How thread dump looks??

it has the thread id and the java.lang.thread.getstackTrace.

complete status of the thread we can know with the help of Thread dump.

we can go to error log in the thread dump and we can observe the thread id and stack trace.

——————————————————————————————————————————————————————

How to find out it is a memory related issue or thread related issue??

Then we have to go and download the heap dump .

——————————————————————————————————————————————————————

Brief explanation of JVM:

whenever you write the code like FOR loop, variables ,If block and methods it will occupy some memory, when you put into a profiler everything you can see in the heap dump, you can see a search graph from that graph you will point out what is the issue.

if you want to see a graphical representation we have to feed to a profiler.

——————————————————————————————————————————————————————

Concept of indexing:

we are creating indexes based on the properties,(In JCR THE REPRESENTATION IS WITH THE HELP OF NODES AND PROPERTIES).(IN REGULAR DATABASES IT IS WITH THE HELP OF ROWS AND COLUMNS).

Examples of indexes:JCR TITLES etc ….

the moment we have this indexes, when i write query i need not iterate through the nodes and properties’, i  can go directly through the indexes.

we can have index for each property.(Note:but we won’t write for every property it is basically related with the query).

——————————————————————————————————————————————————————

When we write a query??
when we want to search ,what you search basically we searched  based on the title,tag and created by  only based on this you can create the index.

Note:AEM 6 comes with the frequently used index store.

that means whatever frequently used properties are there,it already configured for you.

But if you have any custom properties you need to add it to the index.

——————————————————————————————————————————————————————

Goto your CRXDE LITE repo.

In that go to oak index  > we can see all the properties.

just take that property  and paste that in the index and create the index.

——————————————————————————————————————————————————————

If you feel that whenever you create a new page that is not add in to the index.??

then we have to backend in the CRX-quickstart folder > we have the index store .

sometimes we feel  that index store  is outdated, so we need to go to the index manager and do re-index button for that particular property then the updated index will get in to the index store.

what is the difference between the oak and crx2???

we have to get aware of the segments and documents ???

————————————————————————————————————————————————————

log messages will give the list of errors.

——————————————————————————————————————————