How to read a zPulse daily mail: DASD, CPU, space and tape

A zPulse daily mail is the e-mail that a scheduled study sends after each daily run: one mail per family you enable, with the covered period in the subject and at the top, and charts as images in the body. Read it top-down: period first, then the overview charts, then the worst volumes or LPARs, then the history.

The mail is how most people meet zPulse. Engineers use it as a morning check before they open a study; managers use it as their only view. This article explains what the mail is, the six families, how to read the DASD mail and the others, what the mail does not contain, and how to set one up.

What a zPulse daily mail is

A scheduled study is a zPulse study that runs every day at a fixed time on a source folder of reports. A report family is one kind of mail attached to a scheduled study, such as DASD or CPU. You choose which families a scheduled study sends.

Every mail shares the same mechanics:

• Subject. By default, zPulse <family> Report — <study> — <period>, so a mail rule can sort them.
• Covered period. Shown at the top of the mail and in the subject.
• Charts in the body. Images in the mail itself, readable on any client.
• Recipients. To, Cc and Bcc lists, with a subject and body text you can change.
• Your SMTP server. The mail leaves through your own mail server.
• Isolated failures. If one family fails, the others are still sent.

The six mail families

Disk, CPU and DCOLLECT mails come from a scheduled disk study. The three tape mails come from a scheduled tape study.

How to read the DASD mail

The DASD AddInfo mail is the richest. It is organised by time slot of the day.

1. Check the period. Make sure the mail covers the day you expect, and that no source file was missing.
2. Read the I/O chart. It shows the I/O rate over the period, stacked by LPAR if you chose that option. A change in shape is the first clue.
3. Read the response time components. IOSQ, pending, disconnect and connect time, on the same time axis. Look for the component that moved.
4. Read the worst volumes. A table lists the volumes with the highest values. The same names day after day are a pattern; a new name is a lead.
5. Read the filtered sections. You can add sections for an LPAR, a volume serial, a storage group or an LCU, so that critical data has its own view.
6. Compare with history. Charts of the previous days, up to 60 days, show whether today is unusual.

You choose the granularity of the charts: native interval, hour or day.

The schematic below shows where these elements sit in the mail. It is not a zPulse screen: curves, names and figures are example values.

How to read the other mails

• CPU MSU. For each LPAR, three charts: average MSU and summed MSU stacked, and the maximum four-hour rolling average (R4HA) as lines, plus up to 60 days of history. Read the R4HA against the defined capacity first.
• DCOLLECT. Allocated and free space by LPAR as stacked bars, and a scatter of allocated space against the average fragmentation index per volume. An option adds the twenty volumes with the worst fragmentation index.
• Standard report (tape). Active and scratch volumes and capacity for each tape type. It is not sent when no new catalogue file arrived.
• Tape performance. Tape I/O stacked by LPAR with the response time components, with an LPAR filter and the same granularity choices as the DASD mail.
• VEHSTATS. CPU, disk, cache and other charts for each TS7700 cluster, with an option for an aggregated cluster.

What the mail does not contain

The mail is a summary, and some things stay out of it on purpose:

• No attachment. The PDF of charts is written to the report folder of the scheduled study.
• No AI diagnostic. The optional local AI writes its diagnostic of disk studies as text files in the report folder, not in the mail.
• No interactive drill-down. To dig into an interval or a volume, open a manual study on the same reports.
• No history before the first run. The 60-day charts fill up as the scheduled study runs each day.

How to set up a daily mail

In zPulse, a scheduled study is a form with a few blocks: identity and schedule (name, first run date, time), storage and recipients (source folder, report folder, backup folder), report files to write, and one block per mail family. Mails go through the SMTP server set in the general settings, where a test send checks it.

The zPulse server runs as a Windows service, so the study starts on time without anyone logged on. If a long study delays a run, the next check catches it up. The zPulse tools page lists the tools that read the files a scheduled study writes, and our expertise page describes recurring studies commented by an expert.

To see what your own daily mail would show, request a free study and send your RMF or CMF reports.