Alert for messages in RWB with status "wait" or "System error"

[peter_wallner2]

Dear experts,

Is it possible to set up alerts for messages that are stuck in Runtime Workbench?

In general my alerts for stuck messages in SXMB_MONI and adapter errors are working.

Now I am testing messages that have a status "Wait" or after a while a "System error" in RWB but do not trigger an alert.

Why is that? What can I do?

Thank you for any ideas!

Best regards,

Peter

[Former Member · Accepted solution]

Hi Peter,

Runtime workbench is just a tool for monitoring and configurations.

It is a part of JAVA engine already.

So if your alert rule is able to capture the alerts related to adapter engine, then all the errors even if they are because of truncation errors, should be triggered.

-Supriya.

[Former Member · Accepted solution]

Hi peter,

Probabaly the alert rule did not capture the error in the adapter engine.

The filter for sender and receiver interface details in the rule can cause this issue depending on when and how the error occured.

Can you create an alert rule with * for all entries for adapter engine and check?

Also do NOT check the option "suppress multiple alerts from this rule" because it will prevent subsequent alerts from this rule, if there is already one and not yet confirmed.

Further please check the alerts in SALRT table in se16 if its created properly or via transaction ALRTDISP.

Also use the report SXMSALERT_LOGREADER for troubleshooting and checking the alert logs.

regards,

Francis

[peter_wallner2 · Accepted solution]

Hello Francis,

Thank you!

In RWB I created a rule for each adapter we are using and I deleted the general one with adapter type = *

Now I did get the alert once the message in RWB changed from status "wait" to "system error".

I have some messages though I remember that do not change their status from "wait" to "system error" for hours! I would be nice if I could somehow adapt the time span that they change to "system error" faster.

I will see on the board if I can find the answer there!

Best regards,

Peter

[Former Member]

Hi peter,

if you would like to reduce the number of retries and wait time for all channels in the adapter engine then it can be globally set via visual administrator.

choose the service SAP XI Adapter XI,

then adjust the below 2 parameters as you wish, default is 3 and 300,000 milliseconds

xiadapter.inbound.numberRetries.default

xiadapter.inbound.retryInterval.default

Anyway the local settings in RWB - channel monitoring will get the priority if its different from the global one.

hope it helps.

regards,

francis

[peter_wallner2]

Hello Francis,

Thank you Francis for this additional information - I found everything and tried changing the retry settings for the adapter in question in RWB of our test system - it works perfectly!

All questions answered!

Best regards,

Peter

[Former Member]

Hi All,

I am doing the same scenario where I have created two Alert category Alert_cat1 and Alert_cat2, both are similar but one i have created for error in integration engine(Alert_cat1 ) and other for adapter engine(Alert_cat2).

Now for Alert_cat1 i have created a rule and in the setting "where did the error occur" I have given no restriction.

Now for Alert_cat2,as discussed in the above blog I have created many rule each for various adapter.

I have tested it, above mentioned configuration is working for message mapping error (sxmb_moni)

but not for any error in adapter engine for messages in waiting or system erroe status.

Kindly suggest where I am going wrong.

Regards

Shefali S

[Former Member]

Hi Expert,

I am doing the scenario where I have created Alert category named as Alert_cat

Now for Alert_cat i have created a rule for integration engine and in the setting "where did the error occur" I have given "Intergration engine".

Now for same Alert_cat,as discussed in the varios blog I have created many other rules each for various adapter,and in the setting "where did the error occur" I have given "Adapter engine".

I have tested it, above mentioned configuration is working for message mapping error (sxmb_moni)

but not for any error in adapter engine for messages in waiting or system error status.

Kindly suggest where I am going wrong.

Regards

Shefali S

Edited by: shefali s on Dec 20, 2011 10:49 AM

[Former Member]

Hi Peter,

Good to hear that the issue is fixed.

Regarding "wait" by default a message goes in to error in adapter engine and will have 3 retries in a interval of 5 minutes before goes to "error" . during this time the message have the wait status and we get alert only after it has status error after 3 failed retries.

We can locally overwrite this for a specific channel like this.

Go to adapter engine- communication channel monitoring

select the channel that is having the error

Now choose the third tab "settings". you probably sap_xi_administrator_j2ee role to see this tab.

Then you can adjust the delivery attempts for messages.

default is 3 attempts and 5 minutes each in between.

lets say 1 attempt and 3 mintues

Regards,

Francis

[Former Member]

Is it possible to set up alerts for messages that are stuck in Runtime Workbench?

Not possible.

In general my alerts for stuck messages in SXMB_MONI and adapter errors are working.

I am surprised how come you are getting alerts for "stuck messages" in MONI. if they are in error state, then it's working as expected.

Now I am testing messages that have a status "Wait" or after a while a "System error" in RWB but do not trigger an alert.
Why is that? What can I do?

This should work. Check your configuration once again. If possible, share the details of the configuration.

[peter_wallner2]

Hello PI Expert,

In RWB I have created a an alert rule called

"Alert Adapter" and under "Where did the error occur" I set
Adapter Engine --> Adapter Type = *

If there is an error with an adapter, e.g. if an active file adapter does not have a sender agreement I do get an alert.

That alert - as I understand it - does come from RWB, right?

But for example I had a message where a text field was too long and could not be converted from XML to EDIFACT and that error occurred in RWB but I did not get an alert.

How could I configure that?

Thank you again for your help,

Peter