MongoDB Atlas Admin Associate key terms

A group of MongoDB processes holding the same data set, with one primary accepting writes and secondaries applying the oplog. Atlas provisions replica sets by default, giving redundancy and automatic failover without extra configuration.

A topology that partitions collections across multiple shards, each shard being a replica set, with mongos routers and config servers coordinating queries. Atlas uses it when data volume or write throughput outgrows a single replica set.

The node size and storage specification of an Atlas cluster, such as M10, M30 or M60. The tier sets RAM, CPU, IOPS, connection limits and which features are available; M0 and M2/M5 are shared tiers.

A replica set member that can be elected primary during a failover. Distributing electable nodes across regions or availability zones determines where writes may be served and how the cluster survives a zone outage.

A replica set member that stores a full copy of the data and serves reads but can never become primary. Atlas uses these nodes to isolate reporting or analytics traffic from the nodes serving application writes.

A dedicated node, on M10 or higher tiers, targeted by the analytics read preference so reporting queries do not compete with operational traffic. The cluster must be configured with a non-zero analytics node count.

The routing process of an Atlas sharded cluster. Clients connect to mongos, which determines which shard owns the data, dispatches the operation, and merges results returned from multiple shards.

An Atlas setting that raises or lowers a cluster's tier or storage when utilization crosses thresholds you define. It prevents capacity incidents while controlling cost, and can be limited to compute, storage, or both.

A container inside an Atlas organization that groups clusters, database users, network settings and alert configurations. Most day-to-day administration, access control and cost attribution happen at project level.

A credential stored in Atlas that a client uses to authenticate to clusters in a project, carrying roles such as readWrite on a named database. It is distinct from the Atlas UI login used by administrators.

Salted Challenge Response Authentication Mechanism, the username-and-password method MongoDB clients use by default. Atlas supports SCRAM-SHA-1 and SCRAM-SHA-256, and credentials can be rotated or created as temporary users.

A certificate-based login where the client presents a certificate signed by a trusted authority and the server maps its subject to a database user. It requires TLS and suits high-security, password-less environments.

A user-defined bundle of privileges and actions scoped to a database or collection, assigned to Atlas database users. Custom roles implement least privilege when built-in roles grant more access than a workload actually needs.

An Atlas capability that lets database users authenticate through an external OIDC identity provider instead of a static database password. Access is then governed by corporate identity, including group membership and conditional access policies.

Encryption of stored cluster data, backups and snapshots by the underlying cloud provider. Atlas uses provider-managed keys by default and supports customer-managed keys through AWS KMS, Azure Key Vault or GCP KMS.

Encryption performed by the driver before data leaves the client, so the server stores only ciphertext for marked fields. Queryable Encryption, its successor, additionally allows equality queries against encrypted field values.

An Atlas feature that records authentication events and, on dedicated clusters, CRUD and DDL activity into an audit log that can be forwarded to a SIEM. It supports compliance investigations and detection of unauthorized access.

The most privileged project-level Atlas role, able to create and modify clusters, database users, network access entries and alerts. It does not include organization-level capabilities such as billing or managing other projects.

A project-level allowlist of IP addresses and CIDR blocks permitted to reach Atlas clusters. Connections from unlisted addresses are rejected before authentication, so the list must be updated whenever application hosts change.

A direct network link between your cloud provider's virtual network and the Atlas network, carrying traffic over private IP addresses without the public internet. It requires non-overlapping CIDR ranges on both sides.

A private IP address inside your own virtual network that reaches Atlas through a cloud private-link service such as AWS PrivateLink, Azure Private Link or GCP Private Service Connect. It avoids public exposure and CIDR overlap issues.

The URI a driver uses to reach a cluster, containing hostnames, the authentication database, the replica set name and TLS options. Atlas SRV strings begin with mongodb+srv and use DNS to discover cluster members.

Transport encryption between clients and Atlas clusters, enforced by default on all connections. Clients verify the server certificate against a trusted CA and may also present a client certificate for X.509 authentication.

A point-in-time copy of cluster data stored by Atlas, either as a cloud provider snapshot in native storage or as an Atlas-managed snapshot. Snapshots can be restored to the same cluster or a different target cluster.

A backup mode that continuously captures oplog entries alongside periodic snapshots, allowing restoration to any moment inside the retention window. It is available on dedicated clusters and is required for point-in-time restores.

Restoring a cluster to a specific timestamp inside the continuous backup window by replaying oplog entries after a base snapshot. It is the standard recovery path for accidental deletes or a faulty deployment.

A capped collection recording every data-modifying operation so secondaries can replicate it. In Atlas it also feeds continuous backups and change streams, and its retention window limits how far back recovery can reach.

A command-line tool for creating and managing Atlas clusters, database users, network access and backups. It enables scripted, repeatable operations and can emit JSON for use in automation pipelines.

The REST interface for managing Atlas programmatically, covering projects, clusters, database users, alerts and backups. Requests authenticate with API keys and are authorized by the roles assigned to those keys.

An infrastructure-as-code provider that declares Atlas projects, clusters, database users and network settings in Terraform configuration, so environments can be versioned, reviewed and rebuilt consistently.

A public and private key pair used to authenticate to the Atlas Administration API. Keys are scoped to an organization or project and carry roles that limit what the automation is allowed to change.

MongoDB utilities that export a deployment to BSON files and reload them. They suit small or offline migrations and selective restores, but are not ideal for large live clusters where managed migration tools fit better.

A managed service that copies data from an existing MongoDB deployment into an Atlas cluster while the source keeps serving traffic, then performs a short cutover. It minimizes downtime compared with dump-and-restore approaches.

MongoDB's cluster-to-cluster synchronization tool, which continuously copies data and indexes between deployments and can replay changes during cutover. It is used for migrations and for keeping environments in sync.

An Atlas tool that inspects slow queries captured by the profiler and recommends indexes, showing the sample query and a suggested index definition. It is available on dedicated clusters and removes manual log analysis.

The number of documents scanned divided by the number returned by a query. Values far above one signal missing or inefficient indexes, and Atlas can alert when the ratio crosses a configured threshold.

A user-defined alert rule that fires when a metric crosses a threshold or a specific event occurs on a project, cluster or host. Notifications can go to email, webhook, Slack, PagerDuty or similar integrations.

A cluster metric and hard limit describing how many simultaneous client connections the chosen tier allows. Sustained high usage produces connection failures, so teams monitor it and rely on connection pooling.

A recurring window during which Atlas may apply patches or perform infrastructure maintenance, which can include a brief failover. It can be scheduled or deferred so maintenance avoids your peak traffic period.

An Atlas operation that forces the primary to step down so a secondary is elected. Running it verifies that applications use retryable writes and reconnect correctly before a real outage occurs.