Troubleshooting

Configuration Time

Some metrics require a full specified RFC Destination Name like

  • RFC Destination available

  • RFC Latency

The metric RFC Destination available (with logon) is more restrict. It requires a white list. Only destinations collected by the RFC destination extractor and are valid for a remote logon can be chosen. For TCP destinations (Type T) a logon check is not possible.

If the value help does not show any destination at all you should check if the extractor has been scheduled and if it is running successfully.
If you just miss one destination you should check if you can use the logon check from within SM59 of the managed system.

 

 

Reason

The information from F4 Help are usually taken Integration Directory. Here we are reading all Sender and Receiver Agreements and show the related Message Header Attributes for making the data entry more easy. This is just a help, it is still possible to enter the exact values manually (in case the Sender/Receiver Agreements do not cover required Entries). If there are no Sender and Receiver Agreements defined at all, e.g. because ICOs are used in Java-only PI, the F4 Help stays empty.

Solution

In any case the functionality for reading Sender/Receiver Agreements to get Message Header Attributes was not the best solution. But there is no possibility to really get the Message Header Attributes from Integration Directory according tot the defined Scenarios.

The recommendation to use the functionality is:

  • Make sure, the PI Message Statistic extractors are running. E.g. configure an PI related Interface Channel with all '*' values and activate (this will start extraction).
  • Go back to configuration and use the F4 Help to select required entries (since extractors are now running, the value help should be filled).
  • Make sure, all relevant entries are selected (e.g. in case a message fails on the sender system, the receiver header attributes might not be filled yet). You can always compare with the local Message Overview Monitor or central PI Message Overview Monitor (if available).

Reason

In higher SAP Solution Manager Releases: The system is not part of the Technical Scenario for ICMon.

Solution

In higher SAP Solution Manager Releases: Make sure, the system is part of the Technical Scenario for ICMon, that you are currently setting up. 

Reason

In SAP Solution Manager 7.2: Central PI monitoring is prerequisite for defining PI related Channels in ICMon. The Technical Scenario for the PI domain was not created in PI monitoring Setup.

Solution

In SAP Solution Manager 7.2: Go to PI Monitoring Setup and create the Technical Scenario for the PI domain in Step "Define Scope". Follow the PI Monitoring Setup Procedure and finally active PI Monitoring.

Runtime

General Metric Troubleshooting

With report ICMON_TROUBLESHOOT_IS you can check if a channel is configured, activated correctly and if metrics receive in MAI infrestracture and for BI.

Select your interface channel via F4 help. If everything is fine the output should look like:

The traffic light in the first collumn indicates if this step was executed succesfully or not. Further analyis is required:

  1. No MO ID found => Interface channel is not configured or something during auto configuration went wrong; Check SLG1 (Object = E2E_ALERTING) for the user who executed the activation.
  2. No Metric variants found => Interface channel is not configured or something during auto configuration went wrong=> Check SLG1 (Object = E2E_ALERTING) for the user who executed the activation.
  3. No entry in ACE_DB_EVENT found => Metrics are not available (yet) in MAI or might be rejected by gatekeeper.  
  4. Possible problems:
    1. Introscope Push
      • Configuration in MAI does not fit to Introscope EM Push, because Introscope Config.xml could not be sent to EM => Check SLG1 (Object = E2E_ALERTING) for the user who executed the activation.
      • There are no executions/exceptions available in Introscope => Check configured channel and metrics in Introscope investigator
    2. RFC PUSH
      • Problem during collection of metrics => Check MAI servlet
      • Problem when sending the metrics => Check MAI servlet and extractor log 
    3. RFC PULL
      • Extractors are not running or report an error => Check transaction E2E_ALERTING_FWK and enter the SID for which the metrics should be collected.
  5.   No entry in E2EREP_MAPPING found
    1. Is the metric reporting relevant (is the flag set in Activation screen)? Check in Alerting Directory Browser in the metric settings
    2. There is no entry at all in E2EREP_MAPPING for interface channels. Check with MAI_TOOLS → Expert Mode → Configuration → Reporting → Restrict Metrics written to BW (or report SMREP_DPC2BW_EASY_CONFIG) if only selected metrics are written to BW. Make sure ICMOn metrics are written to BW or or set the radio button back to 'All metrics are written to BW'.
    3. Something with mapping for BI did not work. It is possible to display the entries of ACE_DB_EVENT and E2EREP_MAPPING for the context ID (MO ID) when clicking on the table name.

Metric Troubleshooting for PI Channels

Reason

For a better understanding of the analysis path, it is helpful, that the architecture of the current implementation is known.

The PI Message Overview statistics are collected from all PI components of the entire PI domain as part of central PI Monitoring application. These data are stored in SolMan BW.

The ICMon data collector does not collect the same data again. It relies on the data extractors from PI Mon and reuses those data. Therefore the BW is accessed and data (according to the configured selection criteria) are retrieved from the Store of the PI message Overview statistics. This Result is shown in ICMon for PI related Interface Channels.

The PI data extractors for Message Overview statistics are usually scheduled via PI Monitoring Setup. In Solution Manager 7.1 the ICMon Setup will check, if there is PI Monitoring Setup (in fact: if a Tehcnical Scenario of type PI_DOMAIN was created) and then it relies on this PI monitoring Setup tp configure Data colelction. In case the PI mon is not or not yet set up, the ICMon will schedule the PI Data collectors (in solution Manager 7.1). In Solution Manager 7.2 the setup of PI monitoring will be a prerequisite for getting PI related data in ICMon.

Now there could be several reasons, why no data are shown in ICMon for those PI related Interface Channles:

1.  There are no Message Overview statistics data available in PI.

2.  Data are not available in SAP Solution Manager BW.

3.  There are data in BW, but not for the configured Sender and Receiver Header attributes as defined in ICMon Setup.

Solution

Depending on the reason for the issue, there are several things to analyze and correct:

1.  Check, if Messages are available in PI:

  •  
    • Logon to PI Integartion Server (of the related PI Domain) to the /pimon or for older PI releases use /rwb.
    • Open Message Overview Monitor and select the related PI component
    • Check if there are Messages for the Timframe of today or yesterday (or any other timeframe you are using in ICMon)
    • Especially for ABAP Proxies, it is often the case, that there are no Overview statistics available, because of a missing setting in PI. Depending on the PI release, the data extractor in SolMan does not get an error message. Please check the parameter according to SAP note 1031773 (e.g. use report RSXMB_AGG_XPI_ADMIN to check and set the parameter).

2.  Check if Data are available in Solution Manager BW: 

    • Logon to Solution Manager BW.
    • Open either transaction SE16 for table /BI0/ASMD_PIMF00 or transaction LISTCUBE and enter 0SMD_PIMF
    • On Selection Screen enter at lease the PI domain (in SE16 this is field SMD_TSCN) and the timeframes (e.g. SM_PIPTY = Day, SMD_PSTS = <Yesterday>000000 and <Today>000000)
    • If there are no data, try again without timframe entries.
    • If there are data, you can proceed with 3.
    • If there are no data, check the logs of the Extractors: Open Solution Manager Admoinistration Workcenter, Application "Infrastructure" and click in Framework, Extractor Framework.
    • Filter on the Extractors "PI*" and SystemID = <SolMan SID>
    • Usually the root cause for issues here is a not properly executed activitay in Managed System Setup... You can also check the PI monitoring Wiki page for further information regarding Troubleshooting and/or Frequently Asked Questions.

3.  Check the data in Solution Manager BW accorind to configured Sender and Receiver Header attributes

  •  
    • Logon to Solution Manager BW.
    • Open either transaction SE16 for table /BI0/ASMD_PIMF00 or transaction LISTCUBE and enter 0SMD_PIMF
    • On Selection Screen enter at lease the PI domain (in SE16 this is field SMD_TSCN) and the timeframes (e.g. SM_PIPTY = Day, SMD_PSTS = <Yesterday>000000 and <Today>000000)
    • If there are no data, try again without timframe entries, but in standard the collector only takes statistics for today and yesterday into account.
    • If data are shown, you can check the entries in the following fields and compare with your ICMon configuration:
      • SMD_RPRT  =  Reciever Party
      • SMD_RSRV  =  Receiver Component
      • SM_RNAML  =  Receiver Interface (first 60 char) 
      • SM_RNSPL  =  Receiver Namespace (first 60 char)
      • SMD_SPRT  =  Sender Party
      • SMD_SSRV  =  Sender Component
      • SM_SNAML  =  Sender Interface (first 60 char)
      • SM_SNSPL  =  Sender Namespace (first 60 char)
      • Also the Field SM_PISTG  =  "PI Status Group" could be intresting.
    • Configure the ICMon Interface channel accordingly. 

In case you could not find the root casue for the issue or you could not correct it, you can open a customer message on component SV-SMG-MON-IPO-IC. 

 

Metric Troubleshooting for Web Service Channels

One of the following metrics is missing even it is active available in MAI directory:

  • ICMON_IFCHANNEL_WS_EXEC_BELOW_THRESHOLD
  • ICMON_IFCHANNEL_WS_RESPONSE_TIME
  • ICMON_IFCHANNEL_WS_RESPONSE_TIME_INST
  • ICMON_IFCHANNEL_WS_USAGE
  • ICMON_IFCHANNEL_WS_USAGE_INST

To check if there is a problem with metric collection call the MAI servlet and choose the active monitoring agent. To find out the relevant agent for a managed system you can execute report ICMON_TROUBLESHOOT. The relevant agent is displayed and you can also start the MAI servlet for this agent via double-click on the agent.

Reason 1

Is the configuration available?

Click on 'Show Current Configuration' and search for the Metric ID.

When the entry is not available something went wrong during the activation => Check the SLG (Object= E2E_ALERTING) for the timefram when activation has been executed.

Reason 2

Is there an error during collection?

Copy the MetricGroupDefinition ID from previous check, click on 'Show Collection Statistics' and search for the ID. You should see the time stamp and the values retrieved.

Check under 'Last Collection Errors' if you see any error for the '<SID>/abap/com/user'. This would mean the user SMDAGENT_<SID> has authorization issues in the SAP ABAP system.

Metric Troubleshooting for IDoc Channels

IDoc segment filters are not respected => You have configured IDoc (Real-Time Monitoring) or IDoc (Analysis) and want to distinguish the IDocs based on their payload. But still all relevant IDocs contribute to the monitoring result, independent of their payload data. In the job log of BPMon data collection jobs (BPM_DATA_COLLECTION*) on the managed system you might observe "Segment Type not found" messages.

Reason 1:

Insufficient authorization of the user in the TMW / READ destination on the managed system

Solution 1:

Provide corresponding authorizations according to SAP note 784752 (search for term ‘IDoc content').

Reason 2:

Wrong monitoring configuration.

Solution 2:

The input for the IDoc segment filter parameters must have the following syntax: <Segment Name>-<Field Name>. Example: E1EDKA1-PARTN

If you want to filter for qualified segments the syntax <Segment Name>-<Field Name> <Qualifier> must be used. Example: E1EDKA1-PARTN AG

You have configured the IDoc (Real-Time Monitoring) template and face one of the following symptoms:

  • The metrics remain gray in the monitoring application
  • You observe long or even extensive durations of BPMon data collection jobs (BPM_DATA_COLLECTION*) on the managed system whenever IDoc data collection is executed
  • BPMon data collection jobs cancel / memory short dumps are created in program /SSA/EIM
  • You observe extended database locks on table /SSF/BTAB

Reason 1:

BPMon data collection jobs (BPM_DATA_COLLECTION*) don't run as expected in the managed system.

Solution 1:

Check if the user in the TMW destination has the appropriate authorization to release jobs. Check if there's a job hanging in status ‘Released' since quite some time (especially BPM_DATA_COLLECTION_2). Check if relevant ST-A/PI infrastructure SAP notes are applied in the managed system, in particular:

2079462: Advance Corrections in BPMon infrastructure (EXS) delivered with ST-A/PI 01S

2236277: Advance Corrections in BPMon infrastructure (EXS) delivered with ST-A/PI 01S SP

Reason 2:

Too much data has to be processed by the IDoc background routine which accesses the IDoc tables. This often leads to memory dumps. Usually this is due to the fact that IDoc data is selected in a delta approach, i.e. only the IDocs created new since the last collection are relevant for the current data collection run. If the delta to the previous data collection is too big (days, weeks, or even months instead of a few minutes only), the amount of data to be selected is too much. During normal operations the delta won't be too big, but under the following circumstances it can happen that there's a big gap between the previous and the current data collection:

  • BPM_DATA_COLLECTION* jobs got stuck, then resumed after some time (can happen for example after ST-A/PI upgrade)
  • IDoc monitoring was activated for the managed system earlier, then got deactivated. After quite some time it gets activated again. In this case the old timestamp from the previous data collection run is still present on the managed system, which causes the IDoc background routine to select too much data.

Solution 2:

Make sure SAP note 2203583 is implemented on the managed system. The corrections are contained in standard as of ST-A/PI 01S. If the same issue persists follow the instructions in the same note. If this doesn't help either create a customer incident on component SV-SMG-MON-BPM-DCM.

Metric Troubleshooting for File Channels

You have configured the ‘File (ABAP)' monitoring template, and the ‘Creation time of file' metric remains gray.

Reason:

This metric only takes into account files that were created on the current day, after the time specified in configuration parameter ‘Change Time'. If the file was created earlier, the metric doesn't return a measurement.

Solution:

Make sure the ‘Change Time' parameter is aligned with the expected creation time of the file on the managed system. For example, if the file is expected to arrive at 09:00 am, the ‘Change Time' parameter should be set to 08:45 am (to be on the safe side). 

Metric Troubleshooting for qRFC Channels

Reason:

NOSEND doesn't exist as a real status on database (table TRFCQOUT, field QSTATE). Instead the relevant queue entries have status READY with field NOSEND = X.

Solution:

Change the monitoring configuration accordingly (replace NOSEND with READY in the status parameter field).

Monitoring UI

ICMon UI doesn't work correctly in SAP Solution Manager 7.2 SP 04. No scenarios visible on Scenario overview. Topology view grows indefinitely. In Monitoring view only one line is visible even though several channels and metrics are configured.

Solution:

Make sure you implement the latest notes for your SAP_UI version as described in '2090746 - Unified Rendering Notes - Which One To Apply - Instructions And Related Notes'. If this doesn't fix the issue please also implement SAP note '2419950 - SAPUI5 upgrade to version 1.38.17' and all relevant sub-notes. 

You do not get any availability data in the monitoring UI.

Background:

The availability can be measured with a script or for RFC destination starting with Solution Manager 7.1 SP09 via an extractor. The script is usually deployed to the SMD Agent of the source system. If there is no SMD Agent available the SMD Agent of the target system is used. On the SMD Agent the script is not executed by the UXMon executor but by the MAI runtime.

Analysis:

Start in SE38 the report ICMON_TROUBLESHOOT. Enter the SID of the source system (when it is diagnostic relevant otherwise the target system) of the interface channel and execute the report.

 1. If you get the error message No Agent found! ensure that the SMD Agent is installed and running. Via click on the agent name you can open the MAI servlet.Click on 'Show Current Configuration' and search for 'Script'.

2. If the value is not available in the configuration, the configuration has not been pushed to the SMD Agent. In SLG1 you can check to which agents the script has been deployed.

Choose: Object = EEM Subobject = DEPLOY From/to = time when configuration/activation has been executed.

If the script is available you can click on 'Show Collection Statistics' or 'Show Last Collection Errors' to display the result.

3. If you see an exception like Script <script_name> not found you should check if the script really reached the SMD Agent. The file location is under <SMD_AGENT>/applications.config/com.sap.smd.agent.application.eem/rscripts/<script_name>/

If the file is there you should restart the SMD Agent. Either via MMC or via /smd/AgentAdmin → Select Agent and reboot Agent via context menu.

You do not get any exception in the monitoring UI.

Reason:

The DPC Extractor of source ST is not running or returns an error like 'No authorization'. 

Solution:

Check in alerting framwork (transaction E2E_ALERTING_FWK) if the DPC PULL Extractor is running for Extended Context ST. If not schedule DPC Extractor by executing the DPC setup in Basic Configuration.

In case of authorization error upload the role attached note 1899598.