First analysis with the sample trace

In this tutorial you upload the WaitScope synthetic trace through the console, the same trace behind the published sample report, and check each part of the result against that sample. By the end, you will have:

• created an analysis and followed its processing;
• read the total time and the response time profile;
• opened the diagnosis and the plan of the statement that weighed the most.

Before you start

• A console account. During early access, request one by email to contato@waitscope.com. The account comes with a username and a temporary password, which you change at first sign-in.
• An authenticator app on your phone. The console asks for its 6-digit code at every sign-in. When the console offers it, the code can also come by email.
• The synthetic trace saved on your computer as exemplo-10046.trc. It has 36 lines and test timings, with no data from a real database.

Step 1: sign in to the console

1. Open https://app.waitscope.com/ in your browser.
2. Enter User and Password and click Sign in.
3. At first sign-in, register the authenticator app on the Set up two-step verification page and click Confirm app.
4. Keep the recovery codes somewhere other than your computer and click Sign in again.
5. Sign in again with your username and password and enter the app code in Sign-in code.
6. If the console asks, change the temporary password in My password, with 12 characters or more.
7. Check I accept the Terms of Service and have read the Privacy Policy and click Continue.

Result: the Analyses page opens, with the New analysis panel and, under History, No analyses yet.

Step 2: create the analysis

1. In the New analysis panel, click Choose files.
2. Choose exemplo-10046.trc. If you prefer, drag the file to Drag the traces here.
3. Check the summary below the buttons: 1 file selected (1.8 KiB).
4. Click Upload and analyze.

Result: the upload bar reaches the end, File received. Opening the result... appears and the analysis page opens. With a single file, the console does not ask whether the traces come from the same process.

Step 3: follow the processing

The analysis page shows the file name, the status and a progress bar. While queued, the message is Queued. Processing will start shortly. During reading, the percentage of the file already read appears, as in Processing on the server: 40% of the file read. The page asks the server for the status every 2 seconds, without reloading.

You can leave the page: the analysis keeps running on the server, and the History on the Analyses page shows its status, queued, processing, done or error.

Result: when the analysis finishes, the page opens the report by itself. With a trace as small as this one, the report may open right away.

Step 4: read the total time and the profile

1. At the top of the report, under Summary, read the Total time.
2. Scroll down to Response time profile and read the % of total column.
3. Click the name SQL*Net message from client to open the component card.

Result: the Total time is 0.007 s, or 7.0 ms. The profile starts with SQL*Net message from client, at 71.3%, followed by db file sequential read, at 18.5%, and the Total row closes at 100.0. The card explains the component, when it is normal and when it is a concern. Compare with the Summary and the profile of the sample report.

Step 5: open the diagnosis

1. In the In this report index, click Diagnosis.
2. Read the two findings.
3. In the first one, click View in profile.

Result: both findings have medium severity. The first, The session spent most of its time waiting for the application, accounts for 71.3% of total time, and the link leads to the component row in the profile. The second, Statement #1: plan operation 1 accounts for 98.0% of the time, accounts for 24.6% and has the link View in the plan of statement #1. Compare with the Diagnosis in the sample.

The first finding comes before any SQL: most of the time was the session waiting for the next request from the application, and SQL tuning does not reduce that time (session waiting for the application).

Step 6: open the statement and its plan

1. Go back to Diagnosis and, in the second finding, click View in the plan of statement #1.
2. On the statement page, read Share of total time.
3. Scroll down to Execution plan and read the Where the plan spends time panel.

Result: statement #1 is a SELECT with 95.9% of total time. In the plan, operation 1, TABLE ACCESS BY INDEX ROWID T, accounts for 98.0% of the time and 80.0% of the logical reads, and is highlighted in the table. Compare with the detail of statement #1 and with plan operation 1 in the sample.

The published sample is the report downloaded with Download HTML: in it, the detail of each statement opens in place of the report when you click, and Back to the report takes you back; in the console, it opens on its own page. In your analysis and in the sample, the text of statement #1 appears as select * from t where id=***, not id=1 as in the trace: the account comes with both Report privacy options turned on, and the number used as a value becomes *** on the page, in the downloaded report, in the app and in the API. To see the number, uncheck Hide the numbers used as values in the SQL (id = ***) in My account and click Save (hidden literals).

If it does not show up

• The console shows This file was already analyzed on ... Your account already uploaded this trace, and the console opened the existing report. To analyze it again, delete the analysis in History and upload it again (repeated upload).
• The analysis ends with the error the file does not look like an Oracle trace (event 10046). The saved file is not the trace. Download it again from the synthetic trace link.
• The numbers do not match the sample. In the line below the file name, at the top of the report, check that it shows 1.8 KiB and 36 lines: a different file gives a different report.
• Choose at least one trace file appears. No file was chosen, or the chosen file is empty.
• You do not have the authenticator app at hand. On the Sign-in code page, enter one of the recovery codes, in the format xxxxx-xxxxx.

Next steps

• Turn on tracing in the session of the operation you want to understand.
• Find the file on the server and upload it from the browser.
• Read the report section by section, starting with the Summary and the profile, and follow the method to reach the cause.
• Before the period ends, download the report with Download HTML or Download JSON (report retention period).