Meridian · mainframe operations
Watch, act and predict across every LPAR.
Meridian puts the console, the scheduler, the subsystems, the tickets and the batch SLA line for every z/OS LPAR on one screen. Operators act through cards they confirm, and every named batch says in plain words, hours ahead, whether it will meet its SLA.
A free 90-day proof of concept on one LPAR. Meridian runs inside your network; your data never leaves it.
Example data from a lab estate.
Watch
Every LPAR on one screen, with what needs a person at the top.
Dashboards, the console, batch views, health and capacity, and the alerts you set, for every LPAR and every client. What needs a person rises to the top; the rest stays quiet.
- One card per LPAR: live status, WTORs with their age, abends today and every batch in words.
- The console with noise folded, WTORs at the top and the reply box on the line.
- Batch, job abends and SLAs on one page, with the scheduler's own view of each job.
- CICS, IMS, Db2 and Adabas status, spool and initiators, with history per LPAR, read from z/OS and never estimated.
- Alerts on message ids, job names, abend codes, thresholds and SLAs, sent to the channels you choose.
Act
Act from the screen that showed the problem.
Every action is a card: what Meridian saw, what it proposes, what it means for the batch, and the command exactly as it will be sent. One click confirms it, and the outcome lands on the same card.
- WTOR replies on the WTOR line, queued per LPAR so two operators never answer the same WTOR.
- Every scheduler action: restart from a step, force complete, cancel, hold, release, demand, run now, requeue.
- Displays run on a schedule or on Refresh. Anything that changes state goes through a card.
- For a situation seen before, the card shows what was done last time and how often it worked.
Earned autonomy
Meridian does on its own only what you have let it.
Set for each kind of situation and for each client. Move it up when you trust it, and back down at any time. An action Meridian takes on its own is recorded like a clicked one.
Propose only
Meridian raises the card and waits. A person reads it and confirms. Every kind of situation starts here.
Do and tell
Meridian acts, then tells you what it did, in the channel you chose.
Do quietly
Meridian acts without sending a message. The action and its evidence are still recorded, to read and export.
Predict
Every named batch, in plain words, hours before its deadline.
You name the batches that matter and approve their deadlines. Meridian follows each one through the night and says where it stands: On time; On time, but tight; Slightly delayed, within the approved time; or SLA breach incoming.
- The job that decides the batch's end, always named, with how sure Meridian is.
- Each batch opens to its jobs in chain order with their usual and expected ends, its conditions, and your steps for a breach.
- When a job fails, the restart card says whether the batch can still end on time, and by how much.
- Attainment and prediction accuracy by month, per client, and every ended batch marked Done on time, Done late or Missed.
The method
How the prediction is made.
From your own scheduler and your own history, worked out again the moment anything changes. Meridian follows each named batch through the night to its approved deadline.
The chain
Taken from the scheduler's own definitions, CA 7 today. An SME can add or remove jobs, and the screen says what that changes and what it cannot resolve.
Your history
The usual run time of each job, from your own runs, by weekday class, with its spread.
The live estate
The critical path is walked back from the last job. Live subsystem status and initiator counts hold back the jobs that need them.
The words
One of four phrases, with how sure Meridian is. Until there are a few weeks of history, it says it is not sure, in those words.
Builder
Draw what should happen. Meridian runs it and shows every run.
Routine situations, handled the way your team handles them: drawn as a flowchart, run by Meridian, and kept as a timeline of what each step saw, what it did, what came back and how it ended.
- Starts on a console message, a job event, a WTOR, an alert, a ticket assigned or a time.
- Checks and actions in plain forms: subsystem status, job state, replies, commands, scheduler actions, ticket notes, notifications.
- Forks per outcome, and groups for For each LPAR, Repeat and Do these at the same time, laid out by Meridian, never dragged.
- A list view of the same flow for the keyboard and for narrow screens.
- Ready-made flows to copy and change: a ticket for any WTOR or abend, the standard restart, the standard reply.
Tickets
Your ticket tool stays the record. The operator never opens it.
Tickets assigned to the operations group are read, turned into a card with the action they ask for, and written back, so every ticket says what was done and when. ServiceNow today.
- Each ticket with the action Meridian derived from its text, as a card to confirm.
- New tickets for abends and SLA breaches, with the console lines attached.
- Every note, state change and new ticket Meridian wrote, with whether it landed.
- The flows a ticket started, and the settings: the account Meridian uses, how often it reads and what it may do.
Control and evidence
Every client kept apart. Every action on the record.
One install serves a whole company. Meridian runs inside your network, your data never leaves it, and who may do what is set per client.
One install, every client
Each client has its own LPARs, people, roles, alert rules and SLAs, and every screen says which client you are looking at.Four roles, granted per client
Operator, SME, manager and admin. A client's own managers can be given their reports and their audit trail without seeing anyone else's.Four-eyes when a client wants it
Approvals and four-eyes can be switched on per client. They are off by default, because operators are trusted.Evidence for every action
Every reply, command, scheduler action, flow run and ticket write-back, with what Meridian saw, who confirmed it, the time and the outcome. Exportable.Questions
Questions about Meridian.
Short answers. The five parts above have the detail, and the security page covers roles, audit and where your data stays. Or write to hello@hexab.in.
How an engagement runsWhat is Meridian?
Meridian is a mainframe operations platform for z/OS estates, and one of the products Hexabin Infotech makes. It puts the console, the dashboard, the scheduler, the ticket tool and the batch SLA line on one screen, proposes actions as cards an operator confirms, and says in plain words when a batch will miss its SLA, hours before it does.
Is Meridian a monitoring tool or an automation tool?
Both, in one install. Watch is the seeing: dashboards, console, batch views, health and capacity, alerts you set. Act is the doing: WTOR replies, console commands, scheduler actions from cards, flows in Builder, and ticket write-back. Predict adds batch SLA prediction and its reports.
Does Meridian act on its own?
Not until you say so. Every action starts as a card with what Meridian saw, what it proposes, the command exactly as it will be sent, and one Confirm button. For each kind of situation and each client you can move it to do and tell, or do quietly, and back again. Every action is recorded with its evidence either way.
Which schedulers and ticket tools does Meridian work with?
CA 7 and ServiceNow today. Watch works without a scheduler connection; the batch views, scheduler actions and SLA prediction need one. Other schedulers and ticket tools are on the roadmap. Tell us what you run.
How does Meridian predict whether a batch will meet its SLA?
The job chain comes from the scheduler's own definitions. The usual run time of each job comes from your own history, by weekday class, with its spread. The critical path is walked backwards from the last job, live subsystem status and initiator counts hold back the jobs that need them, and the answer is worked out again the moment anything changes. It is one of four phrases: On time; On time, but tight; Slightly delayed, within the approved time; SLA breach incoming.
What does Meridian need from our estate?
Very little. Our engineers install Meridian with your team, inside your network, and it keeps a light footprint on the mainframe. The batch views and SLA prediction need a scheduler account, and scheduler actions need one that may act. Tickets need an account in your ticket tool with read and write on the operations group's tickets. Confident SLA predictions need a few weeks of job history; before that Meridian says it is not sure, in those words.
Where does our data go?
Nowhere outside your network. Meridian runs inside it and your data never leaves it. Every action is recorded with its evidence, and the security page sets out who can see and do what.
How is Meridian priced?
Per LPAR, in two tiers: Watch for seeing everything and Act for doing everything, plus a yearly company fee. There are no per-user fees and volume discounts are built in. The 90-day proof of concept on one LPAR is free. The pricing page has the figures.
See Meridian on your own estate.
Forty-five minutes on a call: Meridian on a live lab estate, then on yours. A free 90-day proof of concept on one LPAR follows if it fits, with our engineers on the call.