While a major revolt is not very different from having a large area revolt in other Paradox games, civil wars are dramatically different. A particularly weak or 
tyrannical ruler can see a significant portion of their characters, armies, and provinces revolt against in them in the name of overthrowing the government.
Triggers
The countdown to a civil war will begin when the 
power base of all disloyal characters as a share of the country's total 
power base is greater than the state's current 
Civil War Threshold. The base threshold is 25%, which is decreased by tyranny and higher country rank and can also be affected by other factors. Larger nations therefore tend to be more prone to civil wars, as they have more powerful characters to handle and need a smaller proportion of them to be disloyal before a civil war can break out. This timer ticks up every month, and if the situation is not brought under control, the civil war will break out after 40 months. During this time, there will be an alert notifying the player that attention should be given to the loyalty of powerful characters immediately. Of course, there are also alerts as soon as a single province or a general becomes disloyal.
It is also possible for civil wars to be triggered outside of the usual civil war counter, most notably as one of the possible outcomes when failing to successfully convict when bringing a powerful character to trial. There are also a number of unique mission and event chains that can trigger civil wars , such as with the invasion of Magos in the 
昔兰尼加反抗军, the overthrow of the File:.png 叙拉古僭主国, and the revolt of the 请检查模板名和国家代码：{{Countries|flag/core|Satrap Coalition}} in the Diadochi missions.
Character loyalty
A character's loyalty to the government is one of their most critical aspects and is important to keep in mind when allocating 
positions and making decisions that affect the loyalty of characters, particularly of those with a large power base. When the loyalty of a character drops below 33, that character is considered disloyal and will begin to actively go against the government, stirring up discontent in other characters, refusing to abandon their positions, armies, or governorships, and generally ignoring any orders from above. If enough powerful characters are disloyal, they will form a bloc to launch a civil war.
There are a large number of factors that affect a character's loyalty, with power base being one of the most important. Powerful characters - especially heads of family - are notable as being both the hardest to keep pleased yet the most dangerous if they become disloyal. Characters of 
scorned families also have a significant loyalty malus, while high tyranny will significantly decrease loyalty across all characters. Monarchies must in particular contend with 
pretenders, particularly if they have a large amount of support (which further decreases their loyalty), while in republics members of factions with 
low approval will also have a significant malus. Certain actions such as revoking positions and 
holdings will also anger characters, if temporarily. However, there are also a large number of ways to increase a character's loyalty temporarily or permanently, such as by making friends, granting a position or holding, giving bribes and free hands, resolving their grievances (if any), and more. A large variety of event options can also temporarily affect loyalty either way. Over the course of the game cumulative loyalty bonuses from 
inventions, 
laws, national ideas, and other permanent modifiers will generally tend to increase loyalty overall.
Generally, it is most important to ensure the loyalty of powerful characters by any means possible, whether by plying them with flattery and bribes or choosing event options to keep them content even at the expense of something or someone else. At the same time, power base should generally be divided up such that no character becomes too powerful and becomes impossible to keep loyal, which for instance means being wary of granting too large and army or governorship to a character that already has a large power base (especially heads of family) or splitting up armies that take up too much of the country's total military, so that its 
commander is not too powerful. If a character's power base is mostly tied to the position they hold (e.g. for governors and commanders), it can sometimes be simpler just to replace them with a more loyal character. Disloyal characters that have a low power base do not pose much of a threat and can generally be ignored, though there is always a chance that they might eventually at some point inherit or come into a position of power.
Cohort loyalty
When led by a commander, every cohort in an army has a monthly chance to become loyal to their commander, based on the total Cohort Loyalty Gain Chance modifiers of the commander, army, and country. The total chance of cohort loyalty is a base of +0.50% per month and is increased by the commander's 
charisma, 
popularity, and 
wealth, by many 
distinctions, and when 
drilling the army, but is reduced to 10% of its normal value when not in battle. Cohort loyalty can also be gained when sacking cities in a war.
When a cohort in a 
legion is loyal to a character, the country pays only 66% of its normal maintenance for it, as the cohort's sponsor will pay part of the unit's upkeep him or herself. A cohort that is loyal to a character also gets a 
+10% discipline modifier. Loyal cohorts, however, always provide power base to the character they are loyal to - significantly more than cohorts that are not loyal to any particular character - and will join their sponsor's side in any civil wars, rather than following their commander, even if the two are different - and even if the character they are loyal to is no longer a commander at all.
Cohort loyalty is generally of little consequence for 
levies, as levied armies are usually not raised long enough for a significant portion of their cohorts to become loyal to their commander, and all cohort loyalty is lost once a levy is disbanded at the end of a war. Instead, cohort loyalty is primarily relevant in legions, where cohorts exist semi-permanently and will be loyal to their sponsor as long as he or she lives. A particularly long-lasting and/or successful commander that has gained the loyalty of many of their cohorts will therefore continue to keep the power base associated with those cohorts even if dismissed, which can make the character quiet powerful and difficult - but important - to appease. Disbanding a loyal cohort in a legion will not completely remove the character's influence but instead converts the cohort into loyal veterans that continue to provide a reduced amount of power base (even though they are no longer part of the army) and can be raised again if their sponsor rebels as part of a civil war.
Mechanics
When the civil war begins, a new dynamic tag is created with all disloyal characters as well as their friends and immediate family members joining the revolt, bringing along all held governorships and holdings as well as commanded 
armies. loyal cohorts, and loyal veterans. 
Disloyal subjects may also side with the civil war revolter, who will additionally start new wars with any nation that the parent country was already at war with. Civil war tags can be recognized by their name, which will always be the adjective of their parent nation followed by "Revolt". The flag will normally be randomly generated, but a few major states have pre-scripted flags for their civil war revolters that reflect an inversion of their normal flags:
- 请检查模板名和国家代码：
{{Countries|flag/core|Roman Revolt}} - 请检查模板名和国家代码：
{{Countries|flag/core|Carthaginian Revolt}} - 请检查模板名和国家代码：
{{Countries|flag/core|Seleukid Revolt}} - 请检查模板名和国家代码：
{{Countries|flag/core|Antigonid Revolt}} - 请检查模板名和国家代码：
{{Countries|flag/core|Egyptian Revolt}} - 请检查模板名和国家代码：
{{Countries|flag/core|Macedonian Revolt}}
Additionally, certain event/mission-triggered civil wars will also have a unique rebellion name.
A civil war is a war-to-the-death, where territories automatically switch owners when they are occupied, and losing a civil war is considered a game over like any other annexation. Peace deals cannot be made during civil wars; the only way to end them is to completely destroy the other side, typically by destroying or decisively defeating all of their armies and taking most to all of their territories. Generally, unless the country is very small civil wars are a large drain on time and resources and should generally be avoided at all costs; equally, a period of civil war is a great time to strike at a now-weakened rival or even a significantly larger power that under normal circumstances would be far too strong to take on.
When a civil war is won, the victorious side will be able to choose what to do with all the characters that sided against them. They can choose to either spare the traitors - who will still permanently carry a stain of dishonour - and regain some of the 
manpower lost from the war, or have all the traitors executed and get a large character loyalty and 
province loyalty boost at the expense of 
integrated culture happiness (as well as losing all the characters who rebelled).
Strategies
- Consolidate the capital region first: Since the
capital region is less prone to rebellion compared to other regions, fully occupying the region helps the ruler with their power base. Conversely, expanding in other regions empowers the regional governor first, before the ruler. - Track the holdings held by family heads: A crucial source of power base for family heads is their holdings, especially if their holdings have many pops. Consider rewarding the family head with a holding with few pops after revoking one with many.
The following thread in Paradox Forum discusses more strategies suggested by players to deal with characters loyalty leading to civil war: