EC2
Uptime SLA is99.95%
within a regionAMIs
are not accessible across Regions, so you need to use the Console or CLI/SDK to copy AMIs between Regions.
- Provides up to
75%
discount Convertible
RI- Have the flexibility to change
family
,OS types
, andtenancies
- Cannot be sold on the Reserved Instance Marketplace
- Have the flexibility to change
- Supports
HTTP
,HTTPS
,TCP
, andSSL
protocols
- Domain registration, DNS routing, and health checking
-
Bucket name restrictions
- Between
3
and63
characters long - Only
lower-case
characters,numbers
,periods
, anddashes
- Must start with a
lowercase
letter ornumber
- Cannot contain underscores, end with a dash, have consecutive periods, or use dashes adjacent to periods
- Cannot be formatted as an IP address (
198.51.100.24
)
- Between
-
ONEZONE_IA
(OneZonal Infrequently Accessed)- Less available and less resilient (
99.50%
availability)
- Less available and less resilient (
-
STANDARD_IA
(Standard Infrequently Accessed)- More available and resilient (
99.9%
availability) - Charges for 128 KB per object (even if the object is less than 128 KB)
- More available and resilient (
-
Regionally redundant except
ONEZONE-IA
-
Glacier
data can be restored within3-5 hours
(can be expedited tominutes
) -
Glacier Deep Archive
data can be restored within12 hours
-
Durability is always
99.999999999%
(11 nines) -
Multipart Upload
- Delivers the ability to begin an upload before you know the final object size
- Delivers quick recovery from network issues
- Delivers the ability to pause and resume object uploads
- Delivers improved throughput
- Provides options for more robust file upload in addition to handling larger files than single part upload
-
Transfer Acceleration
expedites uploads from the Internet to S3 by writing directly to an Edge Location -
The
Pre-Signed
URLs are useful if you want your user/customer to be able to upload a specific object to your bucket, but you don't require them to have AWS security credentials or permissions
- Objects can be restored through
S3
CLI or API
- By default, all accounts are limited to 5 Elastic IP addresses per region
- If the instance is too small, it might not be able to handle the network traffic
- Does not support
edge-to-edge
routing
-
Supports
MySQL
Aurora
PostgreSQL
-
Read Replicas for
MySQL
andMariaDB
support Multi-AZ deployment -
Backup retention period between
0 and 35 days
0
disables automated backupsdefault
is7 days
if you created the DB instance using theConsole
default
is1 day
if you created the DB instance usingRDS API
orAWS CLI
-
When you elect to convert RDS instance from
Single-AZ
toMulti-AZ
, the following happens- A snapshot of your primary instance is taken
- A new standby instance is created in a different Availability Zone
- From the snapshot, synchronous replication is configured between primary and standby instances
-
Multi-AZ RDS deployment will automatically fail-over as a result of
- Loss of availability in primary Availability Zone
- Loss of network connectivity to the Primary
-
Updates are applied to your Read Replica(s) after they occur on the source DB Instance using
Asynchronous Replication
cluster placement group
- Cannot span multiple
AZ
- Can span
VPCs
- Cannot span multiple
spread placement group
cannot use dedicated instances or hosts
EBS
Uptime SLA is99.95%
within a region- SSD volumes must be between
1 GiB - 16 TiB
EBS
volumes is automatically and redundantly stored in multiple physical volumes in the same availability zone at no additional chargegp2
delivers up to 16,000 IOPSInstance Store
- Can provide very high performance but it's ephemeral. Extra precautions would be needed to ensure recovery in the event that the instance itself failed
- Ideal for temporary contents (e.g. buffers, caches, and scratch data)
- Snapshots are constrained to the
Region
in which they were created. To share a snapshot with anotherRegion
, copy the snapshot to thatRegion
EFS
can scale to thousands of concurrent connections. However there are design limits including how manylocks
can be placed against any single file. When you see faults occurring at intervals common binary steps (e.g. 8, 16, 32, 64, etc) always be suspicious that you have hit a limit- The duration of the outage is determined by the number of files changed since the last backup
- The duration of the outage is only related to the initial cataloguing phase
Standard Queue
- Each message will be delivered at least once. This ensures that no message is lost, but leaves you to manage duplicates
- The order that message are processed is loosely sequential, but this cannot be relied on (look for FIFO)
DelaySeconds
delays the message to be visible once it's added to the queuevisibilityTimeout
- The time in which the message becomes hidden while the worker is processing it
- Increasing the visibility timeout will not decrease cost over time.
SSE-S3
uses managed keys and one of the strongest block ciphers available, AES-256, to secure your data at restAWS-KMS
provides an audit trail, so you can see who used your key to access which object and when, as well as view failed attempts to access data from users without permission to decrypt the data.
- Minimum time interval is
1 minute
- Stores metrics for terminated Amazon EC2 instances or deleted Elastic Load Balancers for
15 months
- Whether there is MFA configured on the root account
- Advice on security groups and what ports have unrestricted access
- Maximum number of volumes that can be attached to a VMs during a SMS migration job is
22
- Version numbers are never reused, even for a function that has been deleted and recreated
- You can delete a specific Lambda function version only if there are no aliases dependent on this version
- If you don't specify any version in a
DeleteFunction
request, the entire function including all of its versions and aliases will be deleted - Lambda only publishes a new version if the code hasn't yet been published or if the code has changed when compared against the
$LATEST
version. When you publish additional versions, AWS Lambda assigns a monotonically increasing sequence number for versioning