Baseline data

A baseline is the data from the last several days (usually nine days) aggregated into one “average” or “typical” day. Baselines are necessary for considering the variations in traffic on different days of the week, random anomalies in traffic load, or to compare traffic with a known baseline from a specific point in time.

Baseline data is generated once a day after the arrival of data from the first monitoring interval after 00:10 am (in the background). Baseline data is not averaged over the day within each day and therefore may vary rapidly depending on the time of day – just as monitored data would. Each monitoring interval is assigned the value averaged over the nine-day period for this specific monitoring interval. Requesting baseline data for Yesterday will yield the same results as requesting baseline data for Today, because baseline data for yesterday will still be calculated over the last nine days counting from today.

Note

The daylight saving time (DST) shift generates a discrepancy between current and baseline values. For ten days after the time change, baselines are calculated from a mix of data that originated before the change and after the change. This discrepancy resolves automatically after ten days, when the data used for the baseline calculation originates entirely from after the time change. You can use pinned or manual baselines to mitigate this inconsistency.

Baseline modes

Deprecated

Using a baseline mode other than the regular mode – Daily, rolling over the last N days – is no longer advised.

We no longer offer by default a switch for choosing between baseline data in reports.

The NAM Server enables you to control the way baseline data is calculated. The NAM Server is able to handle day of the week performance patterns (weekly baselines) and to handle static patterns, pinned baselines, and manually imported baselines.

Baseline configuration is done on a stand-alone bases or on the master server. All settings related to baselines are inherited by the slave servers from the master. The baseline calculation is time related: for a specific day of the week or for any day:

Daily

Baselines calculated on the rolling basis (the default is a nine-day rolling average).

Day of the week

Shows how data for the given day of the week compares to previous days. This is similar to a daily baseline but set for a single named day of the week. It is calculated as an average of the last n (4 by default) particular days of the week. For example, the current operation time at Monday, 13:35, compared to an average operation time over the last four Mondays at this hour.

The calculations use either dynamic or static patterns:

Average (typical day)

Changes dynamically every day, based on the data collected for the last several days (defined by the NAM Server data storage period). The “typical” day is a data aggregate for the number of days for which monitoring data resides in the report server database.

Note

NAM Server report server containing several data sources, uses the shortest storage period out of all attached data sources. For example, if the AMD storage period is set to 15, and Dynatrace Synthetic Monitoring – Backbone storage period is set to 10, only the data from the past 10 days will be available.

Pinned baseline

Saved, known state of performance measurements that you consider representative or “good” for a day. For example, a “good” state or a pre-environment change state. You pin a baseline to a defined calendar day.

Manually imported baselines

Similar to pinned baselines except that the representative measurements are imported from an external file. The file is generated by exporting data from DMI. Manual overriding of baselines works in relation to a defined calendar day.

Note that for average daily baselines, data is calculated and aggregated only for the period defined by business days. All baselines and trending data are calculated within the scope of the Daily report generation task.

Working with pinned and average baselines on a single NAM Server

It is possible to enable both average and pinned baseline types. If you do this, each report showing benchmark data will have a switch enabling you to select each type of the baseline data to be presented on reports.

When you create a custom report using DMI baseline-related data views on the Benchmark Data section of the custom report definition, you can use the Baseline source dimension to enable switching between baseline data display on reports.

Example:

Average baseline
Average baseline

Overriding calculated baselines with manually imported values

When baseline calculations are enabled on the NAM Server, you can upload the data to the report server using the values exported from the report database. The data set you provide to the NAM Server overrides the calculated baselines. You cannot use data export for a predefined report on the NAM Server because reports from the main workflow do not have all required dimensions. To obtain expected output, manually define a report using DMI. The process is a multi-step procedure consisting of: creating a custom report, displaying data, and exporting it to the NAM Server using the NAM Console. For more information, see Overriding calculated baselines with manually imported values.

When you work with manually overridden baselines, ensure that at least one baseline type (pinned or average) is enabled. Manually provided baseline values are taken into consideration only after the regular baseline calculation mechanism is triggered and finishes successfully.

Changing baseline configuration

Use the NAM Console to enable or disable generation of baseline data of each type, turn off baseline generation completely, or provide data for overriding automatically calculated baseline data.

  1. Open NAM Console ► Deployment ► Manage devices.
  2. Select NAM Server ► Open configuration for the NAM Server you want to configure.
  3. On the NAM Server configuration screen, choose Baselines and trends.

Disabling baseline generation

Deselect each type of baseline to disable the baseline data generation.

Baseline generation is turned off completely when neither average nor pinned baselines are selected.

By default, Average baselines are enabled.

Note that if you disable baseline generation, you will not be able to use the Network Performance metric, as its calculation depends on the baseline data.

Setting the baseline mode

Baselines are generated either daily or for a particular day of the week.

Select the required type in the Baseline mode section.

The types are mutually exclusive. When you choose a baseline mode based on the day of the week, you are shown the number of days the data is calculated for. This number is dependent on the length of the data retention period on the report server (typically nine days).

Using static patterns for baselines (pinned baselines or manually imported baselines)

To use a saved state of performance measurements from a particular day or period, select Pinned baselines and a calendar day. Make sure that the time you select is within the period for which the NAM Server keeps monitoring data (the data retention period).

Using manually imported baselines is a multi-step procedure consisting of creating a custom report, displaying data, and exporting it to the NAM Server using the NAM Console. For more information, see Overriding calculated baselines with manually imported values.
After you modify any of the existing settings on the screen, click Save to save the new configuration as a draft or directly on the server (publish the configuration instantly).

Advanced settings for baseline and trend aggregation

The average and pinned baselines as well as the trending data aggregates can be calculated for various dimension combinations. You can change the default settings for the available aggregates in Baseline and Trend Configuration in the NAM Console.

Each set of dimensions has a specific set of metrics that are calculated for those dimensions. For example, for software service-server pairs you can measure application performance or the number of slow operations. Some metric values can be compared with baseline data (data from the last several days, usually nine days) aggregated into one “average” or “pinned” day.

The Advanced baselines and trends configuration section lists all dimension combinations for which baselines and trending aggregates are calculated during NAM Server nightly task execution. This means that all selected combinations will be covered automatically by the report server when you display reports containing the data. Accessing the other baselines that are available in DMI may take a while because they are generated on demand for a report that is being defined.

You can select particular combination to enable or disable the calculation of the particular baseline or trend. This calculation is normally performed at midnight, when all the other batch tasks are performed, though it can be performed on user request at any other time.

Caution:

NAM Server performance depends on the number of aggregates for which the baselines and trends are calculated. If you enable additional aggregates, it has a negative effect on NAM Server performance. You can observe the time of generating the baselines and trends on the Capacity Status report available from the Tools ► Diagnostics menu. The Nightly tasks execution time chart shows how long (on average for a selected period of time) it took to execute three tasks: Daily Report Generation (including calculation of trends and baselines), DBDailyMaintenance, and RTMDailyDataStore . The generation time for individual dimension combinations on a single NAM Server can be investigated in Advanced baselines and trends configuration section of baseline configuration screen in the NAM Console.

Overriding calculated baselines with manually imported values

When the baseline calculation mechanism is active on the NAM Server, you can use values from a manually defined DMI report. This report typically presents data matching your criteria and imports them on the NAM Server to override the calculated baseline values.

Exporting data

Note: Do not use a predefined report as a direct source for baseline values. Create a new dedicated report using DMI.

  1. Start a new report definition in DMI.

  2. Select a data view. You use one of the following data views:

    • Software service, operation, and site data (Id: ClientView)
    • Application, Transaction, and Tier Data (Id: Tier)
    • Synthetic and Sequence Transaction Data (Id: Trans)
  3. Required: Add a Time dimension.

    Ensure that:

    • Time is the first column.
    • Data is sorted by Time .
    • The time range is for one day only. Make sure that the start and end time are on the same day. Manual baselines work only for a single day.

    Because the baseline generation engine requires time information, you must always include a time dimension when exporting data for manual generation of baselines. Definitions that do not contain a time definition are not sent to the report server.

  4. Select metrics and dimensions.

    Ensure that the combinations of metrics and dimensions are supported by the manual baselines importing engine. Note that combinations used for manually overridden baselines are a subset of all supported combinations. For more information, see Dimension and metric combinations for manually imported baselines.

  5. Display the report.

    To ensure full coverage of baselines, use a 1 period resolution. If you select a one-hour resolution, manual baselines are only available for the values occurring at the full hour time stamp. In other cases, automatically generated values are used.

  6. Export report data.

    Click Export Data and when the exported data is displayed, save it in a CSV format.

    It is important that the file format be preserved and that it contain all exported fields, including the header part.

Importing data using the NAM Console

  1. Open NAM Console ► Deployment ► Manage devices.
  2. Select NAM Server ► Open configuration for the NAM Server you want to configure.
  3. Select Baselines and trends.
  4. Ensure that at least the Pinned or Average baselines are enabled. Manually imported baselines work only if at least one baseline mechanism is active.
  5. Select Override computed baselines with manual data to enable making changes in this section.
  6. Click Browse and select the CSV file containing baseline data.
  7. Click Save to publish the configuration on the report server or save it as a draft. At this point, you can also revoke the file import by clicking Discard changes instead of clicking Save. During import, only the time (hour and minute) is parsed; the date is not part of the import.

Note that exported values containing null values (represented as “-” on the DMI report) are ignored when the NAM Server receives exported data from the NAM Console. This most commonly occurs when there is a null server IP address for protocols such as ICMP. Because this relates to entities of relatively low importance, your most important values are not affected or distorted.

Dimension and metric combinations for manually imported baselines

Only specific dimensions, metrics, and their combinations can be used for overriding automatically calculated baselines.

Note that all combinations you intend to export must include the Time dimension, in addition to the dimensions listed below. The Time dimension must always be the first element.

Software service, operation, and site data (id: ClientView)

Supported dimensions:

  • Software service
  • Server IP address
  • Operation
  • Link
  • Site

Baseline aggregates are generated for the following dimension combinations on the Software service, operation, and site data data view:

  • Time, Software service
  • Time, Software Service, Server IP address
  • Time, Software Service, Operation
  • Time, Software Service, Link
  • Time, Site
  • Time, Site, Software Service
  • Time, Site, Software Service, Server IP address
  • Time, Site, Software Service, Operation

Metrics Supported for Manual Baseline Calculations for Software Service, Operation, and Site Data:

  • Aborted transactions: tierAbortedTrans
  • Aborts: tierAbortedOper
  • Client RTT: cRtt
  • DNS errors: DnsErr
  • Database errors: DBE1
  • Database warnings: DBE2
  • Errors: diffErrTrans
  • Failures: tierFailedTrans
  • HTTP errors: HTErr
  • Incomplete Responses: RejTrans
  • LDAP errors: LDAPE
  • MQ Errors: mqErr
  • MQ appl. errors: MQAE
  • MS Exchange errors: exchuErrors
  • Network time: tierNetworkTime
  • Operation attributes: AE
  • Operation/Transaction time: tierOperTime
  • Operations/Transactions: tierOper
  • Other time: tierOtherTime
  • RTT measurements: tierRttMeasurements
  • Realized bandwidth: tierRealizedBU
  • Redirect time: RT
  • Responses: tierError
  • SAP GUI errors: sapErr
  • SMTP Errors: SmtpErr
  • SSL errors: sslAlert
  • Server RTT: tierServerRtt
  • Server TCP data packets: tierTcpPacketSvr
  • Server time: tierServerTime
  • Short aborts: saborts
  • Slow operations/transactions: tierSlowOper
  • TCP errors: TCPErr
  • Total redirect time: tierTotRedirectTime
  • Total server time: tierTotalServerTime
  • Total transaction time: tierTotalTransTime
  • Transactional service errors: TuxErrGlobal
  • Forms server error: OFE3
  • HTTP redirect time: HRT
  • HTTP response time (range 1): phttpsvrtimedist1
  • HTTP response time (range 2): phttpsvrtimedist2
  • HTTP response time (range 3): phttpsvrtimedist3
  • HTTP response time (range 4): phttpsvrtimedist4
  • HTTP server image time: HIT
  • HTTP server time: HST
  • Hits: hits
  • Hits (range 1): phttpHitsdist1
  • Hits (range 2): phttpHitsdist2
  • Hits (range 3): phttpHitsdist3
  • Hits (range 4): phttpHitsdist4
  • Hits (started): H
  • Incomplete Responses: RejTrans
  • LDAP error (1): LDAPE1
  • LDAP error (2): LDAPE2
  • LDAP error (3): LDAPE3
  • LDAP error (4): LDAPE4
  • LDAP error (5): LDAPE5
  • Loss rate in: FCL
  • Loss rate out: FSL
  • MQ appl. errors (1): MQAE1
  • MQ appl. errors (2): MQAE2
  • MQ appl. errors (3): MQAE3
  • MQ appl. errors (4): MQAE4
  • MQ appl. errors (5): MQAE5
  • Network performance affected bytes: affTraffic
  • Network performance relevant bytes: relTraffic
  • Number of hits in an aborted transaction: abortHits
  • Operation attributes (1): AE1
  • Operation attributes (2): AE2
  • Operation attributes (3): AE3
  • Operation attributes (4): AE4
  • Operation attributes (5): AE5
  • Operation length: TL
  • Operation load time (range 1): ptrclidist1
  • Operation load time (range 2): ptrclidist2
  • Operation load time (range 3): ptrclidist3
  • Operation load time (range 4): ptrclidist4
  • Operation size (range 1): ptrsizedist1
  • Operation size (range 2): ptrsizedist2
  • Operation size (range 3): ptrsizedist3
  • Operation size (range 4): ptrsizedist4
  • Operation time: transTime
  • Operations: trans
  • Operations with large server time: slowSrvTransactions
  • Oracle Applications error: OFE5
  • Oracle server error: OFE4
  • Orphaned redirects: orphRedirects
  • RTT measurements: RM
  • Realized bandwidth (range 1): pthrputdist1
  • Realized bandwidth (range 2): pthrputdist2
  • Realized bandwidth (range 3): pthrputdist3
  • Realized bandwidth (range 4): pthrputdist4
  • Redirect time: RT
  • Redirect time (range 1): predirtimedist1
  • Redirect time (range 2): predirtimedist2
  • Redirect time (range 3): predirtimedist3
  • Redirect time (range 4): predirtimedist4
  • Request time: RQT
  • SAP GUI errors: sapErr
  • SAP RFC error: SAPRFCErr1
  • SAP RFC error indicator: SAPRFCErr2
  • SSL conn. setup per operation: cST
  • SSL handshake fatal error: sslAlertB
  • SSL handshakes: SSLhd
  • SSL session fatal error: sslAlertA
  • SSL warning: sslAlertN
  • Sampling rate: samplingRate
  • Server ACK RTT: SA
  • Server ACK RTT measurements: SM
  • Server RTT: sRtt
  • Server TCP data packets: SPTCP
  • Server TCP data packets lost: SPL
  • Server bytes: sByte
  • Server loss rate (range 1): pretrsvrdist1
  • Server loss rate (range 2): pretrsvrdist2
  • Server loss rate (range 3): pretrsvrdist3
  • Server loss rate (range 4): pretrsvrdist4
  • Server not responding errors: ClntTmout
  • Server operation size: STrSize
  • Server packets: sPkt
  • Server packets lost (to client): SPLC
  • Server realized bandwidth: srvTput
  • Server session termination errors: AppErr
  • Server time: svrDelay
  • Server time (range 1): ptrsvrdist1
  • Server time (range 2): ptrsvrdist2
  • Server time (range 3): ptrsvrdist3
  • Server time (range 4): ptrsvrdist4
  • Short aborts: saborts
  • Slow operations: slowTrans
  • Slow user sessions: SCS
  • Standalone hits: standalone
  • TCP SYN time: TST
  • Time to abort: abortTime
  • Transaction (or network) throughput: abortTput
  • User sessions: SC
  • Zero window size events: ZWS
  • Dynatrace Hits: dtHits
  • Dynatrace Operations: dtOperations
  • Dynatrace Server errors: dtServerErrors

Application, transaction, and tier data (id: tier)

Supported dimensions:

  • Application
  • Transaction
  • Tier

Baseline aggregates are generated for the following dimension combinations on the Application, transaction, and tier data data view:

  • Time, Application, Transaction, Tier
  • Time, Application, Tier

Metrics Supported for Manual Baseline Calculations for Application, Transaction, and Tier Data:

  • Aborted page Application Delivery Channel Delay: abortADCD
  • Aborted page SSL setup time: abortSslSetupTime
  • Aborted page TCP connect time: abortSetupTime
  • Aborted page http server time: aHTTPSrvTime
  • Aborted page image server time: aImgSrvTime
  • Aborted page redirect time: abortRedirSetupTime
  • Aborted page request size: abortRequestSize
  • Aborted page request time: abortRequestTime
  • Aborted page server delay: aSvrDelay
  • Aborted page size in bytes: abortSize
  • Aborted page transfer time: abortBucketTime
  • Aborts: aborts
  • Attempts: Att
  • Attempts (transport): appl_Att
  • Client ACK RTT: CA
  • Client ACK RTT measurements: CM
  • Client RTT: cRtt
  • Client RTT (range 1): prttclidist1
  • Client RTT (range 2): prttclidist2
  • Client RTT (range 3): prttclidist3
  • Client RTT (range 4): prttclidist4
  • Client TCP data packets: CPTCP
  • Client TCP data packets lost: CPL
  • Client bytes: cByte
  • Client not responding errors: SvrTmout
  • Client operation size: CTrSize
  • Client operations: CT
  • Client packets: cPkt
  • Client packets lost (to client): CPLC
  • Client realized bandwidth: cliTput
  • Connection establishment timeout errors: EstTmt
  • Connection refused errors: conRef
  • Custom metric (1) (cnt): cm1Count
  • Custom metric (1) (sum): cm1Total
  • Custom metric (2) (cnt): cm2Count
  • Custom metric (2) (sum): cm2Total
  • Custom metric (3) (cnt): cm3Count
  • Custom metric (3) (sum): cm3Total
  • Custom metric (4) (cnt): cm4Count
  • Custom metric (4) (sum): cm4Total
  • Custom metric (5) (cnt): cm5Count
  • Custom metric (5) (sum): cm5Total
  • Database errors: DBE1
  • Database warnings: DBE2
  • Delay: DLY
  • Discarded aborts: DiscAborts
  • Failures (TCP): failuresNet
  • Failures (application): FaAppl
  • Failures (transport): appl_FA
  • Forms application error: OFE2
  • Forms client error: OFE1

Synthetic and sequence transaction data (id: trans)

Baseline aggregates are generated for the following dimension combinations on the Synthetic and sequence transaction data data view:

  • Time, Transaction source, Application, Transaction, Analyzer group
  • Time, Transaction source, Application, Transaction, Analyzer group, Site

For each of the dimension combinations, you can enable or disable generation of the following transaction-related metric groups:

Metrics Supported for Manual Baseline Calculations for Synthetic and Sequence Transaction Data:

  • Aborted transactions: abortTrans
  • Application Delivery Channel Delay: adcd
  • Application processing time: appT
  • Attempts (transport): appl_Att
  • Client RTT: cRtt
  • Client bytes: cByte
  • Client packets: cPkt
  • Client response time: clientT
  • Client time: avgIdTime
  • Client time (requests): avgIdTimeAll
  • Custom metric (1) (cnt): cm1Count
  • Custom metric (1) (sum): cm1Total
  • Custom metric (2) (cnt): cm2Count
  • Custom metric (2) (sum): cm2Total
  • Custom metric (3) (cnt): cm3Count
  • Custom metric (3) (sum): cm3Total
  • Custom metric (4) (cnt): cm4Count
  • Custom metric (4) (sum): cm4Total
  • Custom metric (5) (cnt): cm5Count
  • Custom metric (5) (sum): cm5Total
  • Failures (transport): appl_FA
  • HTTP abort error: abortErr
  • HTTP client errors (4xx): HCErr
  • HTTP client errors (category 3): H4xxB3
  • HTTP errors: HTErr
  • HTTP not found errors (404): NH
  • HTTP server errors (5xx): HSErr
  • HTTP server errors (category 1): H5xxB1
  • HTTP server errors (category 2): H5xxB2
  • HTTP timeout error: toutErr
  • HTTP unauthorized errors (401,407): UH
  • Incomplete transaction error: incomplErr
  • Lost pages: lostErr
  • Network time: netDelay
  • No response error: norespErr
  • Operation attributes (1): AE1
  • Operation attributes (2): AE2
  • Operation attributes (3): AE3
  • Operation attributes (4): AE4
  • Operation attributes (5): AE5
  • Operation time: transTime
  • Operations: trans
  • Processing error: procErr
  • RTT measurements: RM
  • Server RTT: sRtt
  • Server bytes: sByte
  • Server packets: sPkt
  • Server time: svrDelay
  • Slow operations: slowTrans

Controlling baseline data generation

The status of trend and baseline generation is available in the diagnostic console under TASKS SCHEDULE STATUS. In case of task failure or timeout, data generation can be launched manually by configuring a manual task using Diagnostic Console.

The system parameters used for trends and baselines generation are in the Advanced Properties Editor :

  • PCS_REPORTING_PERIOD - The number of days used for baseline generation. Only working business days in that range are taken into account.
  • RTM_DATAFILL_FACTOR - The minimum data coverage (in minutes) over a single day that allows for baseline calculation. If a day contains fewer than RTM_DATAFILL_FACTOR minutes of data, a baseline will not be calculated from the data for that day.